Building autonomous systems that move material, navigate the unknown, and walk on two legs.
Built Tesla's autonomous forklift AMR program: SLAM, LiDAR path planning, 3D pallet vision → $2.04M in projected cost savings. Previously led factory automation at Hero MotoCorp. Now finishing my MS in Robotics at Northeastern University with research spanning legged locomotion, deep RL, and optical flow.
- 🎓 MS in Robotics, Northeastern University, Boston (GPA: 3.78), Graduating 2026
- ⚡ Former Manufacturing Equipment Engineer Intern at Tesla Inc., Fremont, CA
- 🤖 Former Robotics Engineer at Hero MotoCorp, India
- 🎯 Open to full-time roles in Robotics, Autonomous Systems, Computer Vision, August 2026
| $2.04M | 83% | 67% | 17% |
|---|---|---|---|
| Cost Savings | Route Complexity Reduction | Pallet Unloading Time Cut | Faster Diagnostics |
| Autonomous forklift AMR at Tesla | Penalty-optimized Theta* planning | Zero-height pallet lifter | RAG-based LLM agent |
- 🦿 Legged Locomotion: Hybrid Zero Dynamics gait design and feedback linearization for bipedal robots
- 🧠 Deep RL for Manipulation: TD-MPC2 policy learning for robotic interception tasks in Genesis sim
- 👁️ Optical Flow: NeuFlow v3 architecture with implicit neural decoders for dense motion estimation
| Project | Description | Tech Stack |
|---|---|---|
| ROBUSTWALKER | Deep RL blind locomotion for Unitree Go1 quadruped. PPO policy in MuJoCo with domain randomization, ~85x GPU speedup, 1.41 m/s learned gait | PPO MuJoCo Stable-Baselines3 PyTorch |
| Gait-Design | HZD-based gait optimization & feedback linearization for 3-link bipedal robot. Machine-precision periodicity (10⁻¹⁵), 1.048 m/s stable walking | MATLAB Nonlinear Control Bézier Optimization |
| VIGIL | Real-time forklift pedestrian safety system. 7 switchable YOLO/RT-DETR models, React dashboard, gRPC streaming, Web Audio alarms | YOLOv8 YOLOv11 RT-DETR FastAPI React gRPC |
| LatentIntercept | TD-MPC2 + Genesis: Franka arm learns to intercept sliding objects (air-hockey RL) with latent dynamics model | TD-MPC2 Genesis PyTorch RL |
| Neuflowv3 | NeuFlow v2 + implicit neural decoder (InfiniDepth-style 3-scale fusion). Best EPE: 3.15px on VKITTI2 | PyTorch Optical Flow Neural Decoder |
| Improved-LLM-A* | Hybrid path planner combining A* with LLM semantic waypoint guidance. ~21.6% reduction in node exploration on 20×20 grids | Python A* LLM Path Planning |
| 3D-Mapping-RTAB-SLAM | RTABMap SLAM for sparse-featured environments (Northeastern tunnel system). ROS 2 Humble, multi-sensor (ZED Mini, RealSense) | RTAB-Map ROS 2 SLAM 3D Mapping |
| RLVR | Reinforcement Learning with Verifiable Rewards. Training LLMs with outcome-based reward signals | Python RL LLM Reward Modeling |
Fremont, CA | Apr 2025 – Dec 2025
- Led autonomous forklift AMR deployment (SLAM + LiDAR + 3D pallet vision) → $2.04M projected savings
- Penalty-optimized Theta* path planning → 83% routing complexity reduction
- Built RAG-based AI LLM agent for machine docs → 17% faster diagnostics
India | 2022 – 2024
- AMR/RGV fleet automation → 23% profitability increase
- Zero-height pallet lifters → 67% pallet unloading time reduction
|
|
|
|
|
| Degree | University | GPA | Key Courses |
|---|---|---|---|
| MS in Robotics (2024 – 2026) | Northeastern University, Boston | 3.78 | Legged Robotics, Robot Sensing & Navigation, Mobile Robotics, Deep Learning for Robotics |
| BTech Mechatronics (2018 – 2022) | SRM Institute, Chennai | 3.81 | Fundamentals of Robotics, Control Systems, Computer Vision |
Robots, SLAM, and path planning. Shipping code, not excuses.
Seeking full-time roles in Robotics & Autonomous Systems | Available August 2026
Ask me about: SLAM, AMR Systems, Legged Locomotion, Deep RL, Sensor Fusion, Path Planning, Computer Vision