I am a Robotics and ML Engineer focused on building intelligent systems that interact with the physical world. My expertise spans Computer Vision, Large Language Models (LLMs), and Robot Operating Systems (ROS2). I have a track record of engineering custom pipelines that reduce computational costs while increasing accuracy.

Currently, I am deep-diving into Robotic Manipulation, specifically working on Inverse Kinematics algorithms to solve complex motion planning challenges.

Robotics
AI & ML
Languages
Full Stack

Innovative Computer Vision pipeline for semantic understanding in cluttered environments.

• Tech: GroundingDINO, CLIP, SAM, Gradio.
• Impact: Improved localization precision by ~38% vs DINO-only baselines.
• Optimization: Implemented semantic filtering & mask-density gating to reduce false detections by ~42%.

Comprehensive AI platform for agricultural intelligence.

• Tech: Custom CV Pipeline, Multilingual LLMs (RAG), Android.
• Performance: Achieved 92%+ accuracy on field crop-disease diagnosis.
• Scale: Deployed tools (Yield Optimizer, Weather Risk Model) used by 50+ farmers in early testing.

Deep learning research implementing core SSL architectures.

• Tech: SimCLR, MAE, MoCo v2, PyTorch.
• Scale: Trained on ImageNet-100 (130k images) under limited compute.
• Results: Reduced pretraining losses by 50–65% within just 5 epochs.

Production-level E-commerce and inventory management.

• Optimization: Reduced manual record-keeping time by ~60%.
• AI Integration: ML-based demand forecasting improved procurement accuracy by ~25%.

• 🏆 National Finalist: Agentic AI Day 2024 by Google Cloud.
• Open Source Developer: Google Developer Student Club (Aug '24 - Present)
  • Optimized ML algorithms for a 20% reduction in rendering time.
• Co-Head, PR Team: Cultural Council DTU
  • Leading operations for Engifest, Delhi’s 2nd largest cultural festival.