Sundeep Teki
  • Home
    • About
  • AI
    • Training >
      • Testimonials
    • Consulting
    • Papers
    • Content
    • Hiring
    • Speaking
    • Course
    • Neuroscience >
      • Speech
      • Time
      • Memory
    • Testimonials
  • Coaching
    • Forward Deployed Engineer
    • Testimonials
  • Advice
  • Blog
  • Contact
    • News
    • Media

Young Worker Despair and Mental Health Crisis in Tech: Data, Root Causes, and Evidence-Based Career Solutions

17/11/2025

Comments

 
​Book a Discovery call​ to discuss 1-1 Coaching to improve Mental Health at work
Picture
Source: https://www.nber.org/papers/w34071
I. Introduction: The Despair Revolution You Haven't Heard About

In July 2025, the National Bureau of Economic Research published a working paper that should alarm everyone in tech. The title is clinical: "Rising Young Worker Despair in the United States."

The findings are significant. Between the early 1990s and now, something fundamental changed in how Americans experience work across their lifespan. For decades, mental health followed a predictable U-shape: you struggled when young, hit a midlife crisis in your 40s, then found contentment in later years. That pattern has vanished. Today, mental despair simply declines with age - not because older workers are struggling less, but because young workers are suffering catastrophically more.
​
The numbers tell a stark story. Among workers aged 18-24, the proportion reporting complete mental despair - defined as 30 out of 30 days with bad mental health - has risen from 3.4% in the 1990s to 8.2% in 2020-2024, a 140% increase. By age 20 in 2023, more than one in ten workers (10.1%) reported being in constant despair. Let that sink in: every tenth 20-year-old colleague you work with is experiencing relentless psychological distress.
This isn't about "Gen Z being soft."

Real wages for young workers have actually improved relative to older workers - from 56.6% of adult wages in 2015 to 60.9% in 2024. Youth unemployment, while higher than adult rates, remains relatively low. The economic fundamentals don't explain what's happening. Something deeper has broken in the relationship between young people and work itself.


For those building careers in AI and technology, this crisis is both personal threat and professional opportunity. Whether you're a student evaluating offers, a professional considering a job change, or a leader building teams, understanding this trend is critical. The same technologies we're developing - monitoring systems, productivity tracking, algorithmic management - may be contributing to the crisis. And the skills we're teaching may be inadequate to protect against it.

In this comprehensive analysis, I'll synthesize macroeconomic research and the future of work for young professionals by combining my experience of working with them across academia, big tech and startups, and coaching 100+ candidates into roles at Apple, Meta, Amazon, LinkedIn, and leading AI startups.

I've seen what protects young workers and what destroys them. More importantly, I've developed frameworks for navigating this landscape that the academic research hasn't yet articulated.


You'll learn:
  • The hidden labor market trends crushing young worker mental health 
  • Why working in tech specifically may amplify these risks
  • The protective factors that separate thriving from suffering young professionals
  • Concrete strategies to build an anti-fragile early career despite systemic pressures
  • Interview questions and red flags to identify toxic setups before accepting offers
  • Portfolio and skill development paths that maximize autonomy and minimize despair risk

This isn't theoretical. The 20-year-olds in despair today were 17 when COVID-19 hit, 14 when social media exploded, and 10 in 2013 when smartphones became ubiquitous. They're arriving in our AI teams with unprecedented psychological burdens. Understanding this isn't optional - it's essential for building sustainable careers and ethical organizations.


II. The Data Revolution: What's Really Happening to Young Workers

2.1 The Age-Despair Relationship Has Fundamentally Inverted
The NBER study, based on the Behavioral Risk Factor Surveillance System (BRFSS) tracking over 10 million Americans from 1993-2024, reveals something unprecedented in the history of work psychology. Using a simple but validated measure - "How many days in the past 30 was your mental health not good?" - researchers identified that those answering "30 days" (complete despair) have fundamentally changed their age distribution:

Historical pattern (1993-2015):
Mental despair formed a U-shape across ages. Young workers at 18-24 had moderate despair (~4-5%), which peaked in middle age (45-54) at around 6-7%, then declined in retirement years. This matched centuries of literary and psychological observation about midlife crisis.

Current pattern (2020-2024):
The U-shape has vanished. Despair now monotonically declines with age, starting at 7-9% for 18-24 year-olds and dropping steadily to 3-4% by age 65+. The inflection point was around 2013-2015, with acceleration during 2016-2019, and another surge in 2020-2024.


2.2 This Is Specifically a Young WORKER Crisis
Here's what makes this finding particularly relevant for career strategy: the age-despair reversal is driven entirely by workers, not by young people in general.

When researchers disaggregated by labor force status, they found:

For WORKERS specifically:
  • Always showed declining despair with age (even in 1990s)
  • BUT the slope has become dramatically steeper
  • Age 18 workers in 2020-2024: ~9% despair
  • Age 18 workers in 1990s: ~3% despair
  • The curve remains downward but shifted massively upward for youth

For STUDENTS:
  • Relatively flat despair across ages
  • Modest increases over time
  • But nowhere near the spike seen in working youth

This labor force disaggregation is crucial. It means: Getting a job - the supposed path to adult stability and identity - has become psychologically catastrophic for young people in a way it wasn't 20 years ago.


2.3 Education: Protective But Not Sufficient
The research reveals stark educational gradients that matter for career planning:


Despair rates in 2020-2024 by education (workers ages 20-24):
  • High school dropouts: ~11-12%
  • High school graduates: ~9-10%
  • Some college: ~7-8%
  • 4+ year college degree: ~3-4%

The 4-year degree provides enormous protection - despair rates comparable to middle-aged workers. This likely reflects both job quality (higher autonomy, better management) and selection effects (those completing college may have better baseline mental health).
However, even college-educated young workers have seen increases. The protective factor is relative, not absolute. A 20-year-old with a 4-year degree in 2023 has roughly the same despair risk as a high school graduate in 2010.

Critical insight for AI careers: College degrees in computer science, data science, or related fields provide significant protection, but the protection comes primarily from the types of jobs accessible, not the credential itself. 


2.4 Gender Patterns: A Complex Picture
The research reveals a surprising gender split:

Among WORKERS:
  • Female workers have higher despair than male workers at all ages
  • The gap is substantial and widening
  • Young women in tech face compounded challenges

Among NON-WORKERS:
  • Male non-workers have higher despair than female non-workers
  • Suggests something specific about male identity tied to employment
  • But also something specifically harmful about women's work experiences

For young women entering AI/tech careers, this is particularly concerning. The field's well-documented issues with sexism, harassment, and lack of representation may be contributing to despair rates that were already elevated. Among 18-20 year old female workers, the serious psychological distress rate (using a different measure from the National Survey on Drug Use and Health) reached 31% by 2021 - nearly one in three.


2.5 The Psychological Distress Data Confirms the Pattern
While the BRFSS uses the "30 days of bad mental health" measure, the National Survey on Drug Use and Health (NSDUH) uses the Kessler-6 scale for serious psychological distress. This independent measure shows identical trends:

Serious psychological distress among workers age 18-20:
  • 2008: 9%
  • 2014: 10%
  • 2017: 15%
  • 2021: 22%
  • 2023: 19%

The convergence across multiple surveys, measurement approaches, and years confirms this is real, not a methodological artifact.


2.6 The Corporate Data Matches Academic Research
Workplace surveys from major employers paint the same picture:

Johns Hopkins University study (1.5M workers at 2,500+ organizations):
  • Well-being scores dropped from 4.21 (2020) to 4.11 (2023) on 5-point scale
  • By 2023, well-being increased linearly with age
  • Ages 18-24: 4.03
  • Ages 55+: 4.28

Conference Board (2025) job satisfaction data:
  • Under 25: only 57.4% satisfied
  • Ages 55+: 72.4% satisfied
  • 15-point satisfaction gap—largest on record

Pew Research Center (2024):
  • Ages 18-29: 43% "extremely/very satisfied" with jobs
  • Ages 65+: 67% "extremely/very satisfied"
  • Ages 18-29: 17% "not at all satisfied"
  • Ages 65+: 6% "not at all satisfied"

Cangrade (2024) "happiness at work" study:
  • Gen Z (born 1997-2012): 26% unhappy at work
  • Millennials/Gen X: ~13% unhappy
  • Baby Boomers: 9% unhappy
The pattern is consistent: young workers are experiencing unprecedented distress, and it's getting worse, not better.


III. The Five Forces Destroying Young Worker Mental Health

3.1 The Job Quality Collapse: Less Control, More Demands
Robert Karasek's 1979 Job Demand-Control Model provides the theoretical framework for understanding what's changed. The model posits that the combination of high job demands with low worker control creates the most toxic work environment for mental health. Modern technological tools have enabled a perfect storm:

Increasing demands:
  • Real-time monitoring of productivity metrics
  • Always-on communication expectations (Slack, Teams, email)
  • Faster iteration cycles and tighter deadlines
  • Reduced "break" times as optimization eliminates "slack" in systems

Decreasing control:
  • Algorithmic task assignment (common in gig work, increasingly in knowledge work)
  • Reduced worker input into scheduling, methods, priorities
  • Remote work paradox: flexibility in location, but often less agency over work itself
  • Junior positions have always had less control, but entry-level autonomy has further declined

In a UK study by Green et al. (2022), researchers documented a "growth in job demands and a reduction in worker job control" over the past two decades. This presumably mirrors US trends. Young workers, entering at the bottom of hierarchies, experience the worst of both dimensions.

For AI/tech specifically:
Many "innovative" tools we build actively reduce worker autonomy:
  • AI-powered productivity monitoring (measuring keystrokes, screen time)
  • Algorithmic management systems that assign tasks without human discretion
  • Performance prediction models that preemptively flag "under-performers"
  • Optimization systems that eliminate buffer time and margin for error
The bitter irony: young AI engineers may be building the very systems that contribute to their own and their peers' despair.


3.2 The Gig Economy and Precarious Contracts
Traditional employment offered a deal: accept limited autonomy in exchange for stability, benefits, and clear career progression. That deal has eroded, especially for young workers entering the labor market.

According to research by Lepanjuuri et al. (2018), gig economy work is "predominantly undertaken by young people." These arrangements create:

Economic precarity:
  • Unpredictable income and hours
  • No benefits, healthcare, or retirement contributions
  • Limited recourse for poor treatment

Psychological precarity:
  • No clear path from gig work to stable employment
  • Constant anxiety about next assignment
  • Inability to plan future (relationships, housing, family)

Career precarity:
  • Gig work often doesn't build traditional credentials
  • Gaps in résumé, difficulty explaining employment history
  • Potential employer bias against non-traditional work

Even young workers in traditional employment face echoes of this precarity through:
  • Increased use of contract-to-hire
  • Longer "probationary periods" before full benefits
  • Performance improvement plans used more aggressively

Maslow's hierarchy of needs places "safety and security" as foundational. When employment no longer provides these, the psychological foundation crumbles.

​
3.3 The Bargaining Power Vacuum
Laura Feiveson from the US Treasury documented the structural shift in worker power in her 2023 report "Labor Unions and the US Economy." The findings are stark:

Union decline disproportionately affects young workers:
  • New entrants join companies with little or no union presence
  • Unable to leverage collective bargaining for better conditions
  • Individual negotiation from position of weakness

Consequences for working conditions:
  • Harder to resist employer-driven changes (monitoring, scheduling, demands)
  • Less recourse when experiencing poor management or harmful conditions
  • Reduced ability to improve terms of employment

The age dimension:
Older workers often in established positions with accumulated social capital within organizations can push back informally. Young workers lack:
  • Reputation and relationships that provide informal protection
  • Knowledge of "how things used to be" to articulate what's changed
  • Credibility to challenge management decisions

This creates an environment where young workers are simultaneously:
  • Subject to the most intensive monitoring and control
  • Least able to resist or modify these conditions
  • Most vulnerable to retaliation if they speak up


3.4 The Social Media Comparison Trap

Multiple researchers point to social media as a key factor, and the timing is compelling:
Timeline:
  • 2007: iPhone launched
  • 2010: Instagram launched
  • 2012-2014: Smartphone penetration reaches majority in US
  • 2013-2015: First signs of age-despair reversal in data

Maurizio Pugno (2024) describes the mechanism: social media creates "material aspirations that are unrealistic and hence frustrating" through constant comparison with idealized versions of others' lives.

For young workers specifically, this operates on multiple levels:
  1. Career comparison: See peers' curated success stories (promotions, launches, awards) without context of their struggles, luck, or full situation
  2. Lifestyle comparison: Observe apparently glamorous lifestyles of influencers, entrepreneurs, or older workers with years of accumulated wealth
  3. Work-life comparison: Remote work during COVID-19 created illusion others have perfect work-from-home setups, while your own feels chaotic
  4. Achievement comparison: In tech especially, cult of the young genius (Zuckerberg, Sam Altman narrative) creates unrealistic expectations

Jean Twenge's research (multiple papers 2017-2024) has documented the mental health decline starting with those who came of age during smartphone era. Those born around 2003-2005, who got smartphones in middle school (2015-2018), are entering the workforce now in 2023-2025 with established patterns of social media-fueled anxiety and depression.

The work connection:
When you're already in distress from your job (high demands, low control, precarious conditions), social media amplifies it by making you feel your suffering is individual failure rather than systemic problem. Everyone else seems fine - must be just you.

​
3.5 The Leisure Quality Revolution
An economic explanation comes from Kopytov, Roussanov, and Taschereau-Dumouchel (2023): technological change has dramatically reduced the price of leisure, particularly for young people.

The mechanism:
  • Gaming devices, streaming services, social media are cheap/free
  • Quality of home entertainment has exploded
  • Cost per hour of leisure enjoyment has plummeted

The implication:
  • Opportunity cost of working has increased
  • Time spent at mediocre job feels more costly when home leisure is so appealing
  • Particularly acute for jobs that are boring, low-autonomy, or poorly compensated

This doesn't mean young people are lazy, it means the value proposition of work has changed. If you're:
  • Working a job with little autonomy
  • Getting paid wages that can't afford a home, relationship, or family
  • Being monitored constantly
  • Having no clear path to improvement

...then spending that time gaming, socializing online, or watching Netflix has higher return on investment.

The feedback loop:
  1. Job sucks → spend more time in leisure
  2. Less invested in work → performance suffers
  3. Lower performance → worse assignments, more monitoring
  4. Job sucks more → cycle continues
For young workers in tech, where much of our work involves building the very technologies that make leisure more appealing, this creates existential tension.


IV. Why AI/Tech Work Carries Unique Risks (And Protections)

4.1 The Autonomy Paradox in Tech Careers

Technology work is often sold to young people as the antidote to traditional employment misery: flexible hours, remote work options, meaningful problems, high compensation. The reality is more complex.

High-autonomy tech roles exist and are protective:
  • Research scientist positions with publication freedom
  • Senior engineer roles with architectural decision rights
  • Product roles with genuine user research input
  • Leadership positions with budget and hiring authority

But young tech workers often enter low-autonomy positions:
  • Junior engineer: assigned tickets, given implementations to code, pull requests heavily scrutinized
  • Associate product manager: doing PM's grunt work without actual decision authority
  • Data analyst: running queries others specify, building dashboards for others' definitions
  • ML engineer: implementing others' model architectures, debugging others' training pipelines

The gap between tech work's promise (innovation, autonomy, impact) and entry-level reality (tickets, micromanagement, surveillance) may create particularly acute disappointment and despair.


4.2 The Monitoring Intensification
Tech companies invented many of the tools now spreading to other industries:

Code monitoring:
  • Commit frequency, lines of code, pull request velocity
  • Code review turnaround times
  • Bug introduction rates, test coverage

Communication monitoring:
  • Slack response times, message volume, "active" status
  • Meeting attendance, video-on compliance
  • Email response latencies

Productivity monitoring:
  • Jira ticket velocity, story point completion
  • Calendar utilization analysis
  • Keyboard/mouse activity tracking (in some orgs)

Performance prediction:
  • ML models predicting flight risk, performance trajectory
  • Algorithmic identification of "low performers"
  • "Data-driven" pip (performance improvement plan) triggering

Young engineers may intellectually appreciate these systems' technical elegance while personally experiencing their psychological harm. You can simultaneously admire the ML architecture of a performance prediction model and hate being subjected to it.


4.3 The Remote Work Double Edge
COVID-19 forced a massive remote work experiment. For young tech workers, outcomes have been mixed:

Positive aspects:
  • Geographic flexibility (live near family, choose low cost-of-living areas)
  • Avoid hostile office environments (harassment, microagressions)
  • Schedule flexibility for medical/mental health appointments
  • Reduced commute stress

Negative aspects:
  • Social isolation, especially for those living alone
  • Loss of informal mentorship (can't absorb knowledge by proximity)
  • Harder to build social capital and reputation
  • Lack of clear work/life boundaries
  • Zoom fatigue and constant surveillance anxiety

The 2024 Johns Hopkins study noted well-being "spiked at the start of the pandemic in 2020 and has since declined as workers have returned to offices and lost some of the flexibility." This suggests the initial relief of escaping toxic office environments was real, but the long-term social isolation and ongoing uncertainty may be worse.

For young workers specifically:
Remote work exacerbates the structural disadvantage of lacking established relationships. Senior engineers can coast on years of built reputation. Junior engineers must build that reputation through a screen, a vastly harder task.


4.4 The AI Skills Protection Factor
Despite these risks, certain AI/ML skills provide substantial protection through creating autonomy and optionality:

High-autonomy skill categories:
  1. Research and experimentation capabilities:
    • Novel architecture design
    • Experiment design and interpretation
    • Theoretical innovation
    • → These skills mean you can self-direct work
  2. End-to-end ownership skills:
    • Full-stack ML (data → model → deployment → monitoring)
    • Product sense (can identify problems worth solving)
    • Communication (can explain and advocate for your work)
    • → These skills mean you can own projects, not just contribute to them
  3. Rare technical capabilities:
    • Cutting-edge model architectures (Transformers, diffusion models, new paradigms)
    • Systems optimization (making models actually deployable)
    • Novel application domains (applying AI to new problems)
    • → These skills provide negotiating leverage
  4. Alternative career paths:
    • Research (academic or industry)
    • Entrepreneurship (technical cofounder value)
    • Consulting (high-end, advisory work)
    • → These skills mean you're not dependent on any single employment path

The protection mechanism:
When you have rare, valuable skills that enable you to either:
  1. Negotiate for better working conditions, or
  2. Exit to alternative opportunities
...you gain autonomy even in entry-level positions. This breaks the high-demand, low-control trap that creates despair.


4.5 The Company Culture Variance
Not all tech companies contribute equally to young worker despair. Based on coaching 100+ candidates and direct experience at multiple organizations, I've observed:

Protective factors in company culture:
  • Explicit mental health support: Not just EAP benefits, but manager training, normalized mental health leave
  • Mentorship structures: Formal programs pairing junior engineers with senior engineers
  • Project ownership path: Clear timeline from support → contributor → owner
  • Manageable on-call: Rotations that respect boundaries, don't create constant alert anxiety
  • Transparent leveling: Understand what's required to advance, how to get there
  • Sustainable pace: 40-50 hour weeks as norm, not exception

Risk factors in company culture:
  • Hero worship: Celebrating all-nighters, weekends, constant availability
  • Stack ranking: Forced curves where someone must be bottom 10%
  • Aggressive PIPs: Using performance improvement plans as stealth firing mechanism
  • Opacity: Decisions made invisibly, criteria for success unclear
  • Constant reorganization: Teams reshuffled every 6-12 months
  • Layoff anxiety: Quarterly speculation about next round of cuts

The interview challenge:
These factors are hard to assess from outside. Section VI will provide specific questions and techniques to evaluate companies before joining.


V. The Systemic Factors You Can't Control (But Need to Understand)

5.1 The Economic Narrative Doesn't Match the Pain

One puzzle in the data: by traditional economic measures, young workers are doing okay or even improving.

Economic improvements:
  • Real wages up 2.4% since 2019 for private sector workers
  • Youth wage ratio to adult workers improved: 56.6% (2015) to 60.9% (2024)
  • Unemployment relatively low (though ~9.7% for 18-24 vs. 3.6% for 25-54)
Yet despair skyrocketed.

This disconnect tells us something crucial: The crisis isn't primarily economic in traditional sense - it's about quality of work experience, sense of agency, and relationship to work itself.

Laura Feiveson at US Treasury articulated this well in her 2024 report:
"Many changes have contributed to an increasing sense of economic fragility among young adults. Young male labor force participation has dropped significantly over the past thirty years, and young male earnings have stagnated, particularly for workers with less education. The relative prices of housing and childcare have risen. Average student debt per person has risen sharply, weighing down household balance sheets and contributing to a delay in household formation. The health of young adults has deteriorated, as seen in increases in social isolation, obesity, and death rates."

Even with improving wages, young workers face:
  • Housing costs: Can't afford home ownership in most markets
  • Student debt: Payments constrain life choices
  • Retirement: Social Security won't exist as currently structured
  • Climate: Future looks objectively worse
  • Inequality: Wealth concentration means mobility illusion

The psychological impact: you can have "good" job by historical standards but feel hopeless because the job doesn't enable the life markers of adulthood (home, family, security) that it would have for previous generations.


5.2 The Work Ethic Shift: Cause or Effect?
Jean Twenge's 2023 analysis of the "Monitoring the Future" survey revealed a startling trend: 18-year-olds saying they'd work overtime to do their best at jobs dropped from 54% (2020) to 36% (2022) - an all-time low in 46 years of data.

Twenge suggests five explanations:
  1. Pandemic burnout
  2. Pandemic reminder that life is more than work
  3. Strong labor market gave workers bargaining power
  4. TikTok normalized "quiet quitting"
  5. Gen Z pessimism about rigged system

Alternative frame:
​This isn't moral failing but rational response to changed incentives. If work no longer delivers:
  • Economic security (wages don't buy homes)
  • Social identity (precarious employment doesn't provide stable identity)
  • Upward mobility (median worker hasn't seen real wage growth in decades)
  • Autonomy and meaning (see all of Section III)
...then why invest deeply in work?

David Graeber's 2019 book "Bullshit Jobs" resonates with many young workers who feel their efforts don't matter, or worse, actively harm the world (ad tech, algorithmic trading, engagement optimization, etc.).

For AI careers:
This creates strategic challenge. The young workers most likely to succeed in AI - those who'll put in years of study, practice, and iteration - are precisely those for whom the deteriorating work contract is most apparent and most distressing.


5.3 The Cumulative Effect: High School to Workforce
The NBER research notes something ominous: "The rise in despair/psychological distress of young workers may well be the consequence of the mental health declines observed when they were high school children going back a decade or more."

The timeline:
  • 20-year-old workers in 2023 were:
    • 17 years old when COVID hit (2020)
    • 14 years old when smartphone use became ubiquitous (2017)
    • 10 years old when Instagram hit critical mass (2013)
  • Youth Risk Behavior Survey (high school students) shows mental health deterioration 2015-2023:
    • Feeling sad/hopeless: 40% girls (2015) → 53% girls (2023)
    • Feeling sad/hopeless: 20% boys (2015) → 28% boys (2023)

The implication:
Young workers aren't entering the workforce with normal psychological baseline and then being broken by work. They're arriving already fragile from adolescence, then encountering work conditions that push them over edge.

For hiring managers and team leads:
The young people joining your AI teams may need more support than previous generations, not because they're weak, but because they've experienced more cumulative psychological damage before ever starting their careers.

For individual young workers:
Understanding this context is empowering. Your struggles aren't personal failure - they're predictable response to unprecedented structural conditions. Self-compassion isn't weakness; it's accurate assessment.


5.4 The Gender Dimension Deepens
The research shows young women in tech face compounded challenges:

Baseline: Women workers have higher despair than men across all ages
Intensified: The gap is larger for young workers
Multiplied: Tech industry adds its own sexism, harassment, representation gaps

Among 18-20 year old female workers, serious psychological distress hit 31% in 2021 - nearly one in three. While this dropped to 23% by 2023, it remains double the rate for male workers (15%).

What this means for young women in AI:
  1. Structural: Face all the same issues as male peers (low control, high demands, precarity) PLUS gender-specific barriers
  2. Social: More likely to experience harassment, discrimination, being ignored in meetings, having ideas attributed to men
  3. Representation: Fewer role models, harder to envision success path, potential impostor syndrome from being numerical minority
  4. Intersection: Women of color face additional dimensions of marginalization

What this means for organizations building AI teams:
  • Can't just hire women and hope for best - must actively create supportive environments
  • Need mentorship structures, sponsorship from senior leaders, zero-tolerance for harassment
  • Must measure and address retention differentials
  • Flexibility and support aren't just nice-to-haves - they're requirements for equitable outcomes


VI. Your Roadmap to Building an Anti-Fragile Early Career

6.1 For Students and Early Career (0-3 years): Foundation Building
The 80/20 for Early Career Mental Health:

1. Prioritize Autonomy Over Prestige
  • Target: Roles where you'll have decision authority within 12 months
  • Example: Small AI startup where you're 3rd engineer >>> Google where you're 1 of 200 on project
  • Why: Prestige doesn't prevent despair; autonomy does
  • How to assess: Ask in interviews: "What decisions will I own in first year?"

2. Build Optionality Through Rare Skills
  • Target: Skills that enable multiple career paths (research, startup, consulting, BigTech)
  • Example: Deep learning fundamentals + systems optimization + communication
  • Why: Optionality = negotiating leverage = autonomy even in entry roles
  • How to develop: Personal projects showcasing end-to-end ownership (see portfolio guide below)

3. Cultivate Relationships Over Efficiency
  • Target: 3-5 genuine mentor relationships (doesn't have to be formal)
  • Example: Regular coffee chats with engineers 3-5 years ahead, not just immediate manager
  • Why: Social capital protects against isolation and provides informal advocacy
  • How to build: Offer value first (help with their side projects, share useful resources), ask thoughtful questions

4. Set Boundaries From Day One
  • Target: 45-hour work week maximum, exceptions require explicit negotiation
  • Example: "I'm working on X tonight" is boundary; "I'm very busy" is not
  • Why: Patterns set in first 90 days are hard to change
  • How to maintain: Track hours, say no to low-value asks, escalate if pressured

5. Develop Alternative Identity to Work
  • Target: Invest 5-10 hours/week in non-work identity (hobby, community, creative pursuit)
  • Example: Music, sports league, volunteering, side business (non-AI), local organizing
  • Why: When work identity fails (layoff, bad manager, etc.), whole self doesn't collapse
  • How to protect: Schedule it like meetings, set boundaries around it

Critical Pitfalls to Avoid:
  • Accepting first offer without comparing culture (You'll spend 2,000+ hours/year there—treat company selection like you'd treat choosing a life partner, not just comparing TC)
  • Optimizing for learning in toxic environment (No amount of technical learning compensates for psychological damage that affects years of career afterward)
  • Staying in bad first job "to avoid job-hopping stigma" (12-18 months is fine - don't stay 3 years in role that's destroying you)
  • Building skills only valued by current employer (If your expertise is "Facebook's internal tools," you're trapped—build portable skills)
  • Neglecting mental health until crisis (Therapy, exercise, sleep, relationships aren't "nice to have" - they're infrastructure for sustainable career)

Portfolio Projects That Build Autonomy:
Instead of just coding what's assigned, build projects demonstrating end-to-end ownership:


Problem identification → Research → Implementation → Deployment → Iteration Example for ML engineer:
  • Identify: "Current ML model for [X] has high false positive rate"
  • Research: Survey literature, test alternative approaches on subset
  • Implement: Build new model with chosen approach
  • Deploy: Package for production, set up monitoring
  • Iterate: Track metrics, communicate results, implement feedback
This demonstrates autonomy and initiative, not just technical chops.


6.2 For Working Professionals (3-10 years): Strategic Positioning
The 80/20 for Mid-Career Protection:

1. Accumulate "Fuck You Money"
  • Target: 12 months expenses in liquid savings
  • Why: Financial runway = ability to leave bad situations = more negotiating power even when staying
  • How: Live below means, aggressive saving even if means smaller house/older car

2. Build Reputation Outside Current Employer
  • Target: Known in broader AI community for specific expertise
  • Example: Papers, blog posts, conference talks, open source contributions, technical Twitter presence
  • Why: Makes you employable elsewhere, which paradoxically makes current employer treat you better
  • How: Dedicate 2-4 hours/week to public work, persist for 18-24 months until compound effects kick in

3. Develop Management and Leadership Skills
  • Target: Ability to lead projects and influence without authority
  • Why: Management track provides different kind of autonomy than individual contributor, and having option is protective
  • How: Volunteer to mentor, lead working groups, run internal talks/workshops

4. Cultivate Strategic Visibility
  • Target: Key decision-makers know your name and your work
  • Example: Brief senior leaders on your projects, contribute to strategy discussions, build relationships with skip-level managers
  • Why: When layoffs or reorganizations hit, visibility = survival
  • How: Communicate proactively, celebrate wins, share insights up the chain

5. Test Alternative Career Paths
  • Target: Explore adjacent opportunities without committing
  • Example: Consulting on side, angel investing, advising startups, teaching, research collaborations
  • Why: Maintains optionality and prevents feeling trapped
  • How: Allocate 5 hours/week, ensure compatible with employment contract

Critical Pitfalls to Avoid:
  • Staying for unvested equity in declining company (Your mental health is worth more than RSUs in company that might not exist)
  • Taking promotion that reduces autonomy (Some "promotions" are traps - more responsibility but less decision authority)
  • Accepting that "this is just how tech is" (Culture varies enormously - don't normalize toxicity)
  • Burning out before asking for help (Flag problems early - easier to fix mild issues than recover from burnout)


6.3 For Senior Leaders (10+ years): Systemic Change
The 80/20 for Leaders:

1. Design for Autonomy at Scale
  • Challenge: How to give junior engineers decision authority while maintaining quality?
  • Framework: Clear domains of ownership with bounded scope, not command-and-control
  • Example: Junior engineer owns "recommendation ranking for mobile web" with clear metrics, full implementation authority

2. Measure and Address Team Mental Health
  • Challenge: Despair is invisible until too late
  • Framework: Regular 1:1s focused on wellbeing, not just project status; anonymous surveys; watch for warning signs
  • Example: Team retrospectives explicitly discuss pace, stress, sustainability

3. Model Healthy Boundaries
  • Challenge: You probably got promoted by working insane hours - now you need to show different path
  • Framework: Visible boundaries (leave at 6pm, take full vacation, unavailable evenings), promote people who work sustainably
  • Example: "I'm off tomorrow for mental health day" in team Slack, showing it's okay

4. Protect Team From Organizational Dysfunction
  • Challenge: Your job includes absorbing chaos so team can focus
  • Framework: Shield from politics, provide context, advocate for resources
  • Example: When reorg happens, communicate quickly and honestly, fight for team's interests

5. Create Paths Beyond Individual Contribution
  • Challenge: Not everyone wants to be principal engineer or manager
  • Framework: Value teaching, mentorship, open source, internal tools as legitimate career paths
  • Example: Promote engineer to senior based on mentorship excellence, not just code output

For organizations seriously addressing young worker despair:
This requires systemic intervention, not individual resilience theater:
  • Mandatory management training on mental health, recognizing distress, creating autonomy
  • Career pathing that's transparent and achievable
  • Compensation that enables life stability (house, family, security)
  • Benefits that include substantial mental health support
  • Culture that celebrates sustainability over heroics
  • Metrics that include team wellbeing alongside technical delivery


VII. Interview Framework: Assessing Company Culture Before You Join

7.1 The Questions to Ask

About autonomy and control:
"Walk me through a recent project. At what point did you [the interviewer] have decision authority vs. needing approval?"
  • Red flag: "Everything needs approval from VP"
  • Green flag: "I owned technical approach, consulted on product direction"

For someone in this role, what decisions would they own outright vs. need to escalate?"
  • Red flag: Vague non-answer or "everything is collaborative" (means no ownership)
  • Green flag: Specific examples of decisions role owns

"How are priorities set for this team? Who decides what to work on?"
  • Red flag: "Roadmap comes from above, we execute"
  • Green flag: "Team has input into roadmap, we balance top-down and bottom-up"

About pace and sustainability:
"What's a typical week look like in terms of hours?"
  • Red flag: "We work hard and play hard" (red flag phrase)
  • Green flag: "Usually 40-45 hours, occasionally more during launch"

"Tell me about the last time you took vacation. Did you check email?"
  • Red flag: Uncomfortable answer or "I caught up on some things"
  • Green flag: "I fully disconnected, team covered for me"

About growth and development:
"How does someone typically progress from this role to next level?"
  • Red flag: "It depends" or no clear answer
  • Green flag: Specific criteria, timeline, examples of people who've done it

"What does mentorship look like here?"
  • Red flag: "Everyone mentors each other" (means no one does)
  • Green flag: Formal program or specific mentor assigned

About mental health and support:
"How does the team handle when someone is struggling with burnout or mental health?"
  • Red flag: Uncomfortable, pivots to EAP benefits
  • Green flag: Specific example of how they've supported someone

About mistakes and failure:
"Tell me about a recent project that failed. What happened?"
  • Red flag: Can't think of one (means not safe to fail) or blames individual
  • Green flag: Describes learning, no finger-pointing


7.2 The Red Flags to Watch For Beyond answers to questions, observe:

During interview:
  • How are you treated? (Respected or talked down to?)
  • Do interviewers seem burned out?
  • Is schedule chaotic? (Interviewers late, disorganized)
  • Do interviewers speak positively about company?

In public information:
  • Glassdoor reviews mentioning overwork, toxicity, poor management
  • LinkedIn showing high turnover (lots of people leaving after 12-18 months)
  • News articles about layoffs, scandals, discrimination lawsuits

During offer process:
  • Pressure to decide quickly
  • Unwillingness to let you talk to potential peers (not just managers)
  • Vague or changing role descriptions
  • Below-market compensation justified as "learning opportunity"
Trust your gut. If something feels off during interviews, it will be worse once you join.


VIII. Conclusion: Building Careers in a Broken System

The research is unambiguous: young workers in America are experiencing a mental health crisis of historic proportions. By age 20, one in ten workers reports complete despair - 30 consecutive days of poor mental health. This isn't weakness. It's a rational response to structural conditions that have made work, particularly entry-level work, psychologically toxic.

The traditional relationship between age and mental wellbeing has inverted. Where previous generations found work provided identity, stability, and a path to adulthood, today's young workers encounter precarity, surveillance, and blocked futures. The promise of technology work—meaningful problems, autonomy, good compensation - often fails to materialize for those starting their careers in AI and tech.

But understanding these systemic forces is empowering, not defeating. When you recognize that:
  • Your struggles aren't personal failure but predictable outcomes of measurable trends
  • Specific, actionable strategies can protect mental health even in broken systems
  • Choices about companies, roles, and skills genuinely matter for outcomes
  • Building autonomy and optionality provides real protection
  • Alternative paths exist beyond the toxic default
...then you can navigate this landscape strategically rather than just endure it.

For students and early-career professionals:
our first job doesn't define your trajectory. Choose companies by culture, not just prestige. Build skills that provide optionality. Set boundaries from day one. Invest in identity beyond work. Leave toxic situations quickly.

For mid-career professionals:
Accumulate financial runway. Build reputation beyond current employer. Develop multiple career paths. Don't mistake promotions for autonomy. Advocate for better conditions.

For leaders:
You have power and responsibility to change systems, not just help individuals cope. Design for autonomy. Measure wellbeing. Model sustainability. Protect teams from dysfunction. Create career paths beyond traditional IC ladder.

The AI revolution is creating unprecedented opportunities alongside these unprecedented challenges. Those who understand both can build extraordinary careers while preserving their mental health. Those who ignore the research will be part of the grim statistics.
You deserve work that doesn't destroy you. The data shows clearly what's broken. The frameworks in this guide show what's possible. The choice is yours.


Coaching for Navigating Young Worker Mental Health in AI Careers

The Young Worker Mental Health Crisis in AI
The crisis documented in this analysis - rising despair among young workers, particularly in high-monitoring, low-autonomy environments - creates both urgent risk and strategic opportunity. As the research reveals, success in early-career AI requires not just technical excellence, but systematic protection of mental health and strategic positioning for autonomy. Self-directed learning works for technical skills, but strategic guidance can mean the difference between thriving and merely surviving.

The Reality Check: The Young Worker Landscape in 2025
  • Mental despair among workers age 18-24 has risen 140% since the 1990s, with 10.1% of 20-year-olds in complete despair by 2023
  • The protective value of education is declining: even college graduates face doubled despair rates compared to a decade ago
  • Job quality has deteriorated faster than compensation has improved, creating gap between economic measures and psychological reality
  • Tech companies lead in deploying monitoring and algorithmic management that reduce worker autonomy - precisely the factor most protective of mental health
  • Gender disparities intensify at young ages, with women in tech facing compounded challenges from both general structural issues and industry-specific sexism
  • Critical window: High school mental health crisis (2015-2023) is now manifesting as workforce crisis (2023-2025), and will intensify

Success Framework: Your 80/20 for Career Mental Health

1. Optimize for Autonomy From Day One
When evaluating opportunities, decision authority matters more than prestige or compensation. A role where you'll own meaningful decisions within 12 months beats a brand-name company where you'll spend years executing others' plans. Autonomy is the single strongest protection against workplace despair.

2. Build Compound Optionality
Every career choice should expand, not narrow, your future options. Rare technical skills, public reputation, financial runway, and alternative career paths create negotiating leverage - which creates autonomy even in junior positions.

3. Strategically Cultivate Social Capital
In remote/hybrid world, visibility and relationships don't happen accidentally. Proactively build mentor network, senior leader relationships, and peer community. These protect against isolation and provide informal advocacy.

4. Set Boundaries as Infrastructure, Not Luxury
Sustainable pace isn't something to establish "once things calm down" - it must be foundational. Patterns set in first 90 days are hard to change. Treat boundaries like technical infrastructure: build them strong from the start.

5. Maintain Identity Beyond Work Role
When work is your only identity, job loss or bad manager becomes existential crisis. Investing in non-work identity isn't self-indulgent - it's strategic resilience that enables risk-taking in career.

Common Pitfalls: What Young AI Professionals Get Wrong
  • Prioritizing company prestige over role autonomy (spending years as small cog in famous machine creates despair even if resume looks good)
  • Staying in toxic first job to avoid "job-hopping stigma" (12-18 months is fine for bad fit - don't sacrifice mental health for outdated employment norms)
  • Building skills only valued by current employer (if your expertise is company-specific internal tools, you're creating dependence, not career capital)
  • Treating mental health as separate from career strategy (your psychological wellbeing IS your career infrastructure - neglecting it guarantees long-term failure)
  • Accepting "this is just how tech is" narrative (culture varies enormously across companies - toxic environments aren't inevitable)

Why AI Career Coaching Makes the Difference
The research reveals a crisis but doesn't provide individualized strategy for navigating it. Understanding that young workers face systematic challenges doesn't automatically translate to knowing which company to join, how to negotiate for autonomy, when to leave a toxic role, or how to build career resilience.

Generic career advice optimizes for traditional metrics (TC, prestige, learning opportunities) without accounting for the mental health implications documented in the research. AI-specific career coaching addresses the unique challenges of entering tech during this crisis:
​
  • Personalized company and role assessment accounting for actual autonomy, not just brand prestige
  • Portfolio development strategies that demonstrate end-to-end ownership and rare skills, creating negotiating leverage
  • Interview question frameworks to assess culture before accepting offers, avoiding toxic environments
  • Compensation and benefits negotiation that includes mental health support, sustainable pace, and autonomy protections
  • Crisis navigation support when you find yourself in bad situation, determining whether to try to fix it or leave strategically
  • Long-term career architecture building toward roles with high autonomy, not just climbing traditional ladder

Who I Am and How I Can Help?
I've coached 100+ candidates into roles at Apple, Google, Meta, Amazon, LinkedIn, and leading AI startups. My approach combines deep technical expertise (40+ research papers, 17+ years across Amazon Alexa AI, Oxford, UCL, high-growth startups) with practical understanding of how career choices impact mental health and long-term trajectories.

Having built AI systems at scale, led teams of 25+ ML engineers, and navigated both Big Tech bureaucracy and startup chaos across US, UK, and Indian ecosystems, I understand the structural forces documented in this research from both sides: as someone who's lived it and someone who's helped others navigate it successfully.

Accelerate Your AI Career While Protecting Your Mental Health
With 17+ years building AI systems at Amazon and research institutions, and coaching 100+ professionals through early career decisions, role transitions, and company selections, I offer 1:1 coaching focused on:

→ Strategic company and role selection that optimizes for autonomy, growth, and mental health - not just TC and prestige
→ Portfolio and skill development paths that build genuine career capital and negotiating leverage, not just company-specific expertise
→ Interview and negotiation frameworks to assess culture before joining and secure roles with meaningful decision authority from day one
→ Crisis navigation and strategic career moves when you find yourself in toxic environments and need concrete path forward

Ready to Build a Sustainable AI Career?
Check out my Coaching website and email me directly at [email protected] with:
  • Your current situation and target roles
  • Specific challenges you're facing with career positioning, company culture, or mental health in tech work
  • Timeline for your next career decision or transition

​I respond personally to every inquiry within 24 hours.

The young worker mental health crisis is real, measurable, and intensifying. But it's not inevitable for your career. With strategic positioning, evidence-based decision-making, and systematic protection of autonomy and wellbeing, you can build an extraordinary career in AI while maintaining your mental health. Let's navigate this landscape together.
References
​[1] Blanchflower, David G. and Alex Bryson, "Rising Young Worker Despair in the United States," NBER Working Paper No. 34071, July 2025, http://www.nber.org/papers/w34071

[2] Twenge, Jean M., A. Bell Cooper, Thomas E. Joiner, Mary E. Duffy, and Sarah G. Binau, "Age, period, and cohort trends in mood disorder indicators and suicide-related outcomes in a nationally representative dataset, 2005–2017," Journal of Abnormal Psychology 128, no. 3 (2019): 185–199

[3] Haidt, Jonathan, The Anxious Generation: How the Great Rewiring of Childhood is Causing an Epidemic of Mental Illness, Penguin Random House, 2024

[4] Feiveson, Laura, "How does the well-being of young adults compare to their parents'?", US Treasury, December 2024, https://home.treasury.gov/news/featured-stories/how-does-the-well-being-of-young-adults-compare-to-their-parents

[5] Smith, R., M. Barton, C. Myers, and M. Erb, "Well-being at Work: U.S. Research Report 2024," Johns Hopkins University, 2024

[6] Conference Board, "Job Satisfaction, 2025," Human Capital Center, 2025

[7] Lin, L., J.M. Horowitz, and R. Fry, "Most Americans feel good about their job security but not their pay," Pew Research Center, December 2024

[8] Green, Francis, Alan Felstead, Duncan Gallie, and Golo Henseke, "Working Still Harder," Industrial and Labor Relations Review 75, no. 2 (2022): 458-487

[9] Karasek, Robert A., "Job Demands, Job Decision Latitude and Mental Strain: Implications for Job Redesign," Administrative Science Quarterly 24, no. 2 (1979): 285-308

[10] Kopytov, Alexandr, Nikolai Roussanov, and Mathieu Taschereau-Dumouchel, "Cheap Thrills: The Price of Leisure and the Global Decline in Work Hours," Journal of Political Economy Macroeconomics 1, no. 1 (2023): 80-118

[11] Pugno, Maurizio, "Does social media harm young people's well-being? A suggestion from economic research," Academia Mental Health and Well-being 2, no. 1 (2025)

[12] Graeber, David, Bullshit Jobs: A Theory, Simon and Schuster, 2019
​

[13] Lepanjuuri, K., R. Wishart, and P. Cornick, "The characteristics of those in the gig economy," Department for Business, Energy and Industrial Strategy, 2018
Comments
comments powered by Disqus
    ★ Checkout my new AI Forward Deployed Engineer Career Guide and 3-month Coaching Accelerator Program ★ ​

    Archives

    November 2025
    August 2025
    July 2025
    June 2025
    May 2025


    Categories

    All
    Advice
    AI Engineering
    AI Research
    AI Skills
    Big Tech
    Career
    India
    Interviewing
    LLMs


    Copyright © 2025, Sundeep Teki
    All rights reserved. No part of these articles may be reproduced, distributed, or transmitted in any form or by any means, including  electronic or mechanical methods, without the prior written permission of the author. 
    ​

    Disclaimer
    This is a personal blog. Any views or opinions represented in this blog are personal and belong solely to the blog owner and do not represent those of people, institutions or organizations that the owner may or may not be associated with in professional or personal capacity, unless explicitly stated.

    RSS Feed

[email protected] 
​​  ​© 2025 | Sundeep Teki
  • Home
    • About
  • AI
    • Training >
      • Testimonials
    • Consulting
    • Papers
    • Content
    • Hiring
    • Speaking
    • Course
    • Neuroscience >
      • Speech
      • Time
      • Memory
    • Testimonials
  • Coaching
    • Forward Deployed Engineer
    • Testimonials
  • Advice
  • Blog
  • Contact
    • News
    • Media