Projects

My projects span reinforcement learning, multi-agent communication, robot coordination, vehicle engineering, and space systems. Each entry links to code, a paper, or an independent reference when one is available.

casRL

Sole creator and principal architect

An open-source research framework that builds RL and MARL experiments from reusable implementations of policies, rollout collection, replay, objectives, optimization, and model updates. casRL implements more than 100 algorithms and variants across online, offline, hybrid, single-agent, and multi-agent learning and sustains 2–4 million environment steps per second during end-to-end single-node training on one CPU and one GPU.

Python PyTorch RL MARL CUDA
Z

Lead researcher

Field-topology feedback based on Gray-code Morton order for coordinating heterogeneous swarms over spatially structured fields. A manuscript is in preparation for ICRA 2027.

MARL Swarm Coordination Spatial Topology
SAR

Creator and first author

A memory-based attention architecture that learns communication and structured reasoning in cooperative and competitive multi-agent tasks. The work appeared at NeurIPS 2020 and includes a multiprocessing TensorFlow training implementation.

MARL Attention Communication TensorFlow
MD

Independent implementation

Reimplemented MD-MADDPG from its published specification when author code was unavailable, including its learned shared-memory communication mechanism.

Python MARL Communication
T-ASE

Research contributor

Developed DeepPaSTL for long-horizon pasture prediction and contributed to a data-synthesis, prediction, planning, and intermittent-deployment pipeline for agricultural robot teams.

Robotics Prediction Planning
LIGO

Research contributor

Contributed to statistical analysis and parameter estimation for GW150914, the binary black-hole merger observed by the Laser Interferometer Gravitational-Wave Observatory.

Statistical Analysis Parameter Estimation
A3C

Distributed Multi-Agent A3C

2018

Research project

Implemented distributed multi-agent asynchronous advantage actor–critic for mixed cooperative–competitive environments using centralized critics and asynchronous actor execution.

RL MARL Distributed Training
HRL

Hierarchical RL for Autonomous Driving

2018

Research project

Investigated hierarchical multi-agent reinforcement learning for joint motion planning and trajectory control in autonomous-driving systems.

Autonomous Driving Hierarchical RL
UAV

Research project

Trained an autonomous flight policy from a large crash dataset using reinforcement learning, ROS, and TensorFlow.

Reinforcement Learning ROS TensorFlow
Team Haya Formula SAE race car during track testing

Suspension and brakes team lead

Led a three-person suspension and brakes team through subsystem design, vehicle integration, and skidpad validation. The car won the Skid Pad event at SUPRA SAE India 2011 against 44 other college teams.

Vehicle Dynamics Suspension Brakes
PE

Power Electronics and Optimization

2012

Undergraduate research

Evaluated neural-network and particle-swarm methods for economic load dispatch and derived state-space averaged transfer functions for non-ideal DC–DC converters with parasitic components.

Power Electronics Neural Networks Optimization