Course | Community | Applied AI | Research | Blogs | Press
Want to crack data science interviews at FAANG & the top tech startups?
- Join my exclusive 4-week cohort-based course featuring mock interview coaching sessions, guest talks by industry leaders, and referrals to hiring companies, this course will transform your career by getting your foot in the door at the top tech companies and startups.
- I will share practical insights from coaching 50+ candidates with offers at Apple, Meta, Amazon, LinkedIn, Twitter in the US.
- The 1st Cohort concluded (Feb. 2022) with an NPS of 80 (rare for any EdTech product).
Community
Join 200+ AI/ML learners at my AI-First Discord server to learn & engage with like-minded AI folks and grow your network
Applied AI
- I led Conversational AI R&D at Swiggy, India’s largest food ordering and delivery platform where I developed novel Deep Learning technologies for NLP and Speech. In NLP, I worked on Natural Language Understanding for chatbot, Intent as well as Sentiment Classification models. In Voice AI, I led research on Speech recognition for Hinglish, Voice sentiment analysis and led the team to publish Swiggy’s first AI paper.
- I worked as a Research Scientist at Amazon Alexa AI in Seattle, USA, where I worked on advancing Alexa's Speech Recognition and Natural Language Understanding AI capabilities. I developed state-of-the-art end-to-end deep learning models for speech recognition using sequence models and distributed deep learning, and deployed high impact NLP models to detect and address sensitive and offensive conversations with Alexa for multiple languages using neural machine translation.
Research
Deep learning for NLU & NLP
- NLU modeling using weak supervision to decode intent in code-mixed chat
- Sentiment analysis to identify negative feelings, emotions and opinions in chat
- Language identification in code-mixed chat
- Predicting Social Media Escalations based on chat input
- Classification of Products into multiple categories based on text inputs
- Named Entity Recognition of various entities in chat conversations
- NLP for detecting sensitive and offensive chat interactions
- Trained deep learning NLP models to detect sensitive content in human-assistant interactions
- Deployed NLP models for non-English languages using Neural Machine Translation and Text Generation techniques
Deep learning for Speech
- Developed state-of-the-art end-to-end deep learning ASR models based on sequence-to-sequence models
- Significantly improved ASR model training times using Horovod for distributed training
- Trained ASR models on >20000 hours of data on AWS GPU instances
- ASR modeling for "Hinglish", i.e. Hindi interspersed with English common amongst Indian speakers
- Spoken Language Identification for low-resource Indian languages code-mixed with English
- Speaker Diarization modeling, i.e. "who spoke when" for code-mixed speech
- Audio Sentiment Analysis to understand emotions expressed in speech
Privacy-preserving deep learning
- Homomorphic encryption, federated learning and its applications for private machine and deep learning
Papers
Gupta A, Sukumaran R, John K, Teki S (2021)
Hostility Detection and Covid-19 Fake News Detection in Social Media
CONSTRAINT Workshop, AAAI 2021 (non-archival) [download]
Vijjali R, Potluri P, Iyer S, Teki S (2020)
Two Stage Transformer model for Covid-19 Fake News Detection and Fact Checking
NLP for Internet Freedom Workshop co-located at COLING 2020 [download] [data]
Rangan P*, Teki S* (2020) [*Equal contribution as first authors]
Exploiting Spectral Augmentation for Code-Switched Spoken Language Identification
1st Workshop on Spoken Language Technologies for Multilingual Communities, INTERSPEECH 2020 [download]
Brahma AK, Potluri P, Kanapaneni M, Prabhu S, Teki S (2021)
Identification of Food Quality Descriptors in Customer Chat Conversations using Named Entity Recognition
Accepted at CODS-COMAD 2021 Research Track. [download]
Teki S (2019)
Internationalization of NLP Models for Sensitive Content Detection in Alexa Utterances
Amazon Machine Learning Conference
Gupta A, Sukumaran R, John K, Teki S (2021)
Hostility Detection and Covid-19 Fake News Detection in Social Media
CONSTRAINT Workshop, AAAI 2021 (non-archival) [download]
Vijjali R, Potluri P, Iyer S, Teki S (2020)
Two Stage Transformer model for Covid-19 Fake News Detection and Fact Checking
NLP for Internet Freedom Workshop co-located at COLING 2020 [download] [data]
Rangan P*, Teki S* (2020) [*Equal contribution as first authors]
Exploiting Spectral Augmentation for Code-Switched Spoken Language Identification
1st Workshop on Spoken Language Technologies for Multilingual Communities, INTERSPEECH 2020 [download]
Brahma AK, Potluri P, Kanapaneni M, Prabhu S, Teki S (2021)
Identification of Food Quality Descriptors in Customer Chat Conversations using Named Entity Recognition
Accepted at CODS-COMAD 2021 Research Track. [download]
Teki S (2019)
Internationalization of NLP Models for Sensitive Content Detection in Alexa Utterances
Amazon Machine Learning Conference
Blogs
- Choosing a Data Governance Framework for your Organization [Domino Data Lab]
- Surefire Ways to Identify Data Drift [Unbox.ai]
- Data Labeling and Relabeling in Data Science [Unbox.ai]
- Understanding and Measuring Data Quality [Unbox.ai]
- Benefits of FAANG companies for Data Science & ML roles [BecomingHuman.ai]
- Why Data Democratization is important to your business? [Transform]
- When is the right time to migrate to Kubernetes? [ContainIQ]
- How to ensure Data Quality through Governance [Transform]
- Data Labeling: The Unsung Hero Combating Data Drift [LabelStud.io]
- ML Engineer vs Data Scientist [Neptune.ai]
- Knowledge Distillation: Principles, Algorithms & Applications [Neptune.ai]
- Best Practices for Improving Machine Learning Models [Neptune.ai]
- Developing AI/ML Projects for Business - Best Practices [Neptune.ai]
- How to build AI Teams that Deliver? [Neptune.ai]
- Why Corporate AI Projects Fail? Part 1/4 [BecomingHuman.ai]
- Why Corporate AI Projects Fail? Part 2/4 [BecomingHuman.ai]
- TLDR: AI for Text Summarization and Generation of TLDRs [Towards Data Science]
- Covid or just a Cough? AI for Detecting Covid-19 from Cough Sounds [KD Nuggets]
- Reskilling India for an AI-First Economy [BusinessWorld]
- Fact-checking Covid-19 Fake News [Towards Data Science]
- AI-enabled Conversations with Analytics Tables [Towards Data Science]
- My recent work on Covid-19 Fake News Detection [1, 2] is profiled in IIT Madras' Science and Technology Magazine, Shaastra.
- Here's my recent interview for Haptik.ai on AI R&D, NLU, Voice Tech, Startup Innovation and Culture - https://www.conversation.ai/interview/ai-research-and-innovation-that-matters
- Here's my talk at Inflection Point Ventures on 'AI for Startups'
| AI for Startups @IPV |
- Here's my talk at Metvy on 'Careers in AI' for high-school and undergraduate students.
| 2021_AI & Careers @Metvy |