| Profile |
| AI/ML Research Engineer with a PhD in Machine Learning and 8+ years building production AI systems at scale. Expert in deep learning, NLP, computer vision, and MLOps. Published 12 peer-reviewed papers with 800+ citations. Led AI teams at Google and Meta deploying models to 100M+ users. Passionate about safe, explainable, and impactful AI. |
| Technical Stack |
Frameworks
PyTorch · TensorFlow · JAX · Keras
|
NLP / LLMs
Transformers · LangChain · OpenAI API · Llama
|
MLOps
MLflow · Kubeflow · SageMaker · Vertex AI
|
Languages
Python · R · C++ · CUDA · SQL
|
Infra
AWS · GCP · Docker · Kubernetes · Spark
|
Databases
PostgreSQL · MongoDB · Pinecone · Weaviate
|
|
| AI Projects |
LLM Fine-tuning Pipeline
Built scalable RLHF + DPO fine-tuning pipeline for 7B–70B parameter LLMs. Reduced alignment training cost by 60%.
60% cost reduction
|
Real-time Fraud Detection
Graph neural network fraud detection system processing 2M transactions/day with 99.7% precision at Meta Pay.
99.7% precision · 2M tx/day
|
Multimodal Search Engine
CLIP-based multimodal search for Google Shopping. Improved product relevance by 28% vs. text-only search.
+28% relevance
|
AutoML Platform
Internal NAS + HPO platform reducing model development cycle from 3 months to 2 weeks for 40+ Google teams.
3mo → 2wk cycle
|
|
| Industry Experience |
| Senior AI Engineer |
Jan 2023 – Present |
| OpenAI · San Francisco, CA |
| › |
Research and development on GPT-4 and ChatGPT alignment and safety systems |
| › |
Developed novel RLHF sampling strategies improving instruction-following by 22% |
| › |
Led cross-functional AI safety red-teaming exercises with 30+ researchers |
| › |
Published 3 papers on constitutional AI and reward model design |
|
|
| ML Research Scientist |
Jun 2019 – Dec 2022 |
| Meta AI Research (FAIR) · Menlo Park, CA |
| › |
Developed graph neural network models for content ranking serving 3B+ users |
| › |
Co-invented fraud detection system reducing financial losses by $120M annually |
| › |
Mentored 5 PhD interns resulting in 4 ICML/NeurIPS paper acceptances |
| › |
Patented 2 novel ML architectures (US Patent 11,234,567 and 11,345,678) |
|
|
| Machine Learning Engineer |
Aug 2016 – May 2019 |
| Google Brain · Mountain View, CA |
| › |
Built multimodal embedding models for Google Shopping (Image + Text) |
| › |
Deployed AutoML platform adopted by 40+ product teams across Google |
| › |
Contributed to TensorFlow Extended (TFX) open-source ML pipeline framework |
|
|
| Selected Publications |
| [1] |
"Reward Hacking in RLHF: Detection and Mitigation" — NeurIPS 2023 (Oral · 142 citations) |
|
| [2] |
"Graph Neural Networks for Real-time Fraud Detection" — ICML 2022 (Best Paper Award) |
|
| [3] |
"Multimodal Contrastive Pretraining for E-commerce Search" — ACL 2021 |
|
| [4] |
"Neural Architecture Search at Scale with Population-Based Training" — ICLR 2020 |
|
|
Education
| PhD · Machine Learning / Computer Vision |
2016 |
| Stanford University, AI Lab (SAIL) |
NSF Graduate Research Fellowship · Best Thesis Award
Advisor: Prof. Fei-Fei Li |
| Master of Science · Computer Science |
2013 |
| MIT · GPA 5.0/5.0 |
| EECS Department Award · Valedictorian |
| Bachelor of Technology · Computer Science |
2011 |
| IIT Bombay · GPA 9.7/10 |
| Institute Gold Medal |
|
Certifications
| TensorFlow Developer Certificate |
2022 |
| Google |
| AWS Machine Learning Specialty |
2021 |
| Amazon Web Services |
| Deep Learning Specialization |
2020 |
| Coursera / deeplearning.ai |
Expertise
Deep Learning
Natural Language Processing
Computer Vision
Reinforcement Learning
LLMs / GPT
MLOps
Distributed Training
Data Engineering
Research
Python
PyTorch
TensorFlow
|
|