I’m a Postdoctoral Researcher at Meta Superintelligence Labs in London, working with Yoram Bachrach on AI Research Agents (AIRA). Before that, I completed two research internships at Meta: at FAIR London with Nicolas Carion (SAM 3, AI research agents), and at Reality Labs Research in Redmond with Ruta Desai and Akshara Rai (EgoTV, ICCV 2023).
My PhD, supervised by Luc De Raedt and Pedro Zuidberg Dos Martires, focused on LLM agents that use search to reason and make decisions.
I completed my Master’s in AI at the Indian Institute of Science (IISc) in 2019, supervised by Ambedkar Dukkipati. I also worked on natural language understanding at Amazon Alexa-AI, and on multi-agent reinforcement learning as a Research Associate in IISc’s Statistics and Machine Learning Group.
news
Sep 24, 2026
Paper accepted at NeurIPS 2026: AIRA2: Overcoming Bottlenecks in AI Research Agents.
Sep 05, 2026
🏅 Gold medal for AIRA3! Our AI research agent finished 8th of ~4,000 teams in NVIDIA’s live Kaggle competition on fine-tuning a 30B Nemotron model.
May 01, 2026
Paper accepted at ICML 2026: Training AI Co-Scientists Using Rubric Rewards.
Apr 01, 2026
Mar 20, 2026
Paper accepted at GECCO 2026 COvolve: Adversarial Co-Evolution of LLM-Generated Policies and Environments via Two-Player Zero-Sum Game.
Jan 25, 2026
Paper accepted at ICLR 2026 SAM 3: Segment Anything with Concepts.
Oct 17, 2025
Successfully defended my PhD! 
Thesis.
Thesis. Sep 25, 2025
2 papers accepted at NeurIPS 2025: AI Research Agents as Spotlight, LexiCon in DB Track.
Jul 08, 2025
Paper accepted at COLM 2025: Have Large Language Models Learned to Reason? A Characterization via 3-SAT
May 01, 2025
Paper accepted at IJCAI 2025 (Demo Track): Combining Code Generating Large Language Models and Self-Play to Iteratively Refine Strategies in Games.
Feb 26, 2025
Invited Talk at UCSB NLP Group on A Computational Theory Perspective on Reasoning
Jan 22, 2025
2 Papers accepted at ICLR 2025 REvolve: Reward Evolution with Large Language Models using Human Feedback (main track) & Can Large Language Models Reason? A Characterization via 3-SAT.
Jan 07, 2025
Paper accepted at HRI 2025 (Late Breaking Report): Evaluating Efficiency and Engagement in Scripted and LLM-Enhanced Human-Robot Interactions.
Nov 03, 2024
Awarded Top (8%) Reviewer for NeurIPS 2024. Twice in 3 years of reviewing for NeurIPS.
Aug 26, 2024
Starting as a Research Scientist Intern at FAIR, London under Nicolas Carion.
Aug 20, 2024
Paper accepted at LLMRA@RoMan 2024: Bidirectional Intent Communication: A Role for Large Foundation Models.
May 02, 2024
EgoTV dataset is now available on Meta AI datasets.
Apr 19, 2024
Talk on SayCanPay: Planning with LLMs at TAILOR.
Mar 01, 2024
Invited talk on Planning & Robotics with LLMs at Bosch Robotics Seminar. Link to Slides.
Jan 11, 2024
Demoing Robot Planning and Manipulation framework at WASP Winter Conference 2024, Norrköping, Sweden .
Dec 09, 2023
Paper accepted at AAAI 2024: SayCanPay: Heuristic Planning with Large Language Models using Learnable Domain Knowledge.
Jul 14, 2023
Paper accepted at ICCV 2023: EgoTV: Egocentric Task Verification from Natural Language Task Descriptions.
Jun 06, 2023
Paper accepted at ECML-PKDD 2023: Deep Explainable Relational Reinforcement Learning: A Neuro-Symbolic Approach.
May 12, 2023
Presented EgoTV: Egocentric Task Verification from Natural Language Task Descriptions at the Embodied AI Seminar in Meta AI (FAIR).
Feb 23, 2023
Position talk on A (Large) Language Model is ALL YOU NEED at the Centre for Applied Autonomous Sensor Systems (AASS), Örebro University
Jan 10, 2023
Presented poster on Deep Explainable Relational Reinforcement Learning: A Neuro-Symbolic Approach at the WASP Winter Conference 2023, Norrköping, Sweden
Jul 25, 2022
Started as Research Science Intern @ Meta Reality Labs, Redmond, Washington under the mentorship of Ruta Desai, Akshara Rai, and Nitin Kamra
May 12, 2022
Chaired Session on Reasoning and Logic @ AAMAS’22
academic service
Session Chair & Program Committee
AAMAS'22 (Session Chair) NeurIPS'22–26 ICML'23–25 ICLR'24–25 KR'24 EACL'23 TMLR