Chitwan Saharia
AI Research Scientist at Meta
Old Toronto, Ontario, United States
Summary
👤
Senior🎓
Top SchoolChitwan Saharia is an AI research scientist with nine years of experience specializing in deep reinforcement learning and NLP, currently working at Meta MSL TBD Labs after founding Ideogram. He has a strong research-to-production trajectory through roles at Google (AI Resident to Senior Research Scientist) and hands-on work at MILA under Yoshua Bengio on hierarchical dialogue and grounded language learning. His open-source contributions include optimizing BabyAI’s curriculum learner—improving batch sampling, early stopping, and TensorBoard instrumentation—to make RL training workflows more robust and observable. Comfortable bridging research and engineering, he has repeatedly shipped and debugged complex learning systems and led technical teams in startup and large-company settings. Based in Old Toronto, he brings both academic rigor from IIT Bombay and practical product-facing experience in deployed AI systems.
9 years of coding experience
5 years of employment as a software developer
Indian Institute of Technology Bombay
English, Hindi
Stackoverflow
31reputation
9kreached
1answer
1question
Github Skills (7)
imitation-learning10
machine-learning10
python10
tensorboard9
nlp8
pdb6
uniform-distribution6
Programming languages (4)
Github contributions (5)
mila-iqia/babyai
Jun 2018 - Oct 2018
BabyAI platform. A testbed for training agents to understand and execute language commands.
Role in this project:
ML Engineer
ML EngineerContributions:77 commits, 6 comments, 5 issues in 3 months
Contributions summary:Chitwan primarily focused on modifications and debugging within the curriculum learner, a core component of the BabyAI platform. Their contributions involved adjustments to the curriculum learner's parameters and early stopping mechanisms, indicating an effort to optimize the training process. The code changes included modifications to the batch sampler and the integration of tensorboard logging, suggesting improvements in the training workflow and performance monitoring of reinforcement learning environments.
nlpagentsmachine-learningimitation-learningnlp-machine-learning
iterative-refinement/iterative-refinement.github.io
Jun 2021 - Nov 2021
Contributions:54 commits, 54 pushes in 5 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.