CV

Education

PhD, Aeronautics and Astronautics, MIT, (Expected 2025)
SM, Aeronautics and Astronautics, MIT, 2022
B.Tech, Electrical Engineering, IIT Madras, 2020

Theses

Learning-based Scheduling
- S.M. Thesis, Massachusetts Institute of Technology, 2022
- Advisor: Prof. Hamsa Balakrishnan
Reinforcement Learning for Improving Object Detection
- Undergraduate Thesis, Indian Institute of Technology Madras, 2020
- Advisor: Prof. Balaraman Ravindran

Work experience

DINaMo Lab, MIT, Graduate Research Assistant - Fall 2020-Present
- Advisor: Prof. Hamsa Balakrishnan
  - Developed and deployed a language model-based architecture for multi-agent embodied robotic tasks resulting in 30% improvement in task completion.
  - Designed a graph neural network-based architecture for scaling multi-agent reinforcement learning algorithms in limited-information scenarios resulting in 1.9× improvement in success rates.
  - Exploring LLMs for creating interpretable communication protocols for multi-agent reinforcement learning.
  - Developed a hybrid solution combining reinforcement learning (RL) and integer programming to optimize crew schedules, achieving 33% - 48% fewer disruptions compared to the baseline formulation.
MERL, Cambridge, MA, Research Scientist Intern, Summer 2024
- Advisor: Dr. Anoop Cherian
  - Developed a multimodal language model for aligning visual, aural, and language modalities for autonomous navigation in embodied robotic environments leading to 21% improvement in prediction accuracy of navigation actions.
  - Implemented a data-distributed parallel RL model to train a conversational embodied robotic agent to include inputs from the trained multimodal language model.
MERL, Cambridge, MA, Research Scientist Intern, Summer 2023
- Advisor: Dr. Abraham Vinod
  - Developed a data-driven bi-level approach incorporating multi-armed bandits and integer programming to enhance multi-agent environmental monitoring strategies
  - Devised a real-time implementable graph-based heuristic planner, significantly improving solution speed
  - Obtained anytime guarantees and upper bounds on computing time as well as task completion time
TCS Research and Innovation Lab, Mumbai, India, Research Intern, Summer 2019
- Advisor: Dr. Harshad Khadilkar
  - Co-developed a heuristic + reinforcement learning-based method for the online version of 3D-bin packing problem to improve the packing efficiency by 3% over heuristic-based methods along with speeding up prediction by a factor of 6.
  - Experimented with a behavioural cloning + heuristics model to achieve 85% average packing efficiency
RISE Lab, IIT Madras, UnderFall 2018-2019: Undergraduate Research Assistant
- Advisor: Prof. Balaraman Ravindran
  - Utilized RL to enhance digital transformations on images, optimizing object detection performance in a pre-trained network.
  - Applied graph neural networks for molecule property prediction, achieving a notable ROC-AUC of $0.807$ on the Pseudomonas dataset. Ranked 13th globally in the MIT AI Cures Challenge. Leaderboard
Daimler AG R&D, Sindelfingen, Germany, Research Intern, Summer 2018
- Advisor: Dr. Hannes Gorniaczyk
  - Analyzed the impact of on-board vehicle camera parameters, including Shutter Speed and Voltage Gains, on the performance of pre-trained object detection neural networks.
  - Developed a performance matrix to identify the optimal combination of shutter speed and voltage gain, maximizing the F1-score for a pre-trained object detection network.

Skills

Multi-Agent Reinforcement Learning
Graph Learning
Foundational Models
Distributed Parallel Computing
Slurm
Python
- PyTorch
- TensorFlow
C, C++
ROS (Robot Operating Software)

Talks

NASA ULI Safe Aviation Autonomy Seminar. “Scalable Multi-Agent Reinforcement Learning through Intelligent Information Aggregation”
Tata Consultancy Services Research and Innovation Labs. –”–

Service

Conference Reviewing

AAAI (2021, 2024)
CVPR (2024)
IROS (2024)
IFAC (2024)
NeurIPS (2024)
ICLR (2024)
ACL (2024)

Journal Reviewing

IEEE Transactions on Circuits and Systems for Video Technology (2023)
Complex & Intelligent Systems (CIS)
Information Science (IS)
IEEE Robotics and Automation Letters (RAL)
Journal of Guidance, Control, and Dynamics (JDCD)

Workshop Reviewing

The 4th Workshop on Mathematical Reasoning and AI @NeurIPS (2024)
NeurIPS 2024 Workshop on Multimodal Algorithmic Reasoning @NeurIPS (2024)
Robotic Tasks and How to Specify Them? @RSS (2024)

Workshop Organisation

Coordination and Cooperation in Multi-Agent Reinforcement Learning (CoCoMARL) @RLC 2024

Mentoring

A list of UROPs+MEngs I have mentored:

Wenqi Ding (EECS, S.B., MIT) 2022, 2024
Vittal Thirumalai (EECS, S.B.+ M.Eng.), 2024
Jackson Zhang (EECS, S.B.+ M.Eng.), 2024
Adelmo Morrison Orozco (EECS + Math, S.B.) 2024
Marina Ten Have (EECS + AI, S.B.) 2024
Darren Chen (EECS, S.B.) 2024
Daniel Liu (EECS, M.Eng.) 2021-2022
Kenneth Choi (EECS, S.B.) 2022
Carson Smith (EECS, M.Eng) 2021
Laura Peralta (Electrical and Electronics, B.E.; Hampton University) 2021
Akila Sarvanan (AeroAstro, S.B.; MIT) 2021
Simran Pabla (AeroAstro, M.Eng.; MIT) 2021

Honors and Awards

R&D 100 Awards - Software and Services Category (2023)
USAF Analytics Excellence Award for the Space Training and Readiness Command (STARCOM) (2022)
HULT Prize: Chennai Regional WInner and Singapore Regional Finalist (2017-2018)
Ranked among the top 1% in the National Standard Examination in Physics (Physics Olympiad) (2016)