|
|
"Engineering at the intersection of silicon and intelligence — building autonomous quadrupeds,
air-gapped LLM inference engines, and real-time computer vision deployed directly on edge hardware."
- 🤖 Robotics & AI Engineer at BIT Bangalore (2023–2027), currently conducting research at Computational Data Sciences (CDS), IISc Bangalore.
- ⚡ Core Focus: Machine learning, deep learning, computer vision, quadruped robotics, kinematics, and local LLM systems.
- 🛠️ Engineering Philosophy: Bare-metal optimization, low-latency ZeroMQ inter-process IPC, and shipping production-ready AI models.
|
Dept. of Computational & Data Sciences Benchmarked 17 ML/DL models (CNN, BiLSTM, CNN-LSTM) on MXene piezoresistive sensor data. Feature-engineered 21 time/frequency domain parameters with 5-fold cross-validation in collaboration with AGH University, Poland. |
Robotics & Edge AI Engineered custom IK solvers & 12-state FSM gait controllers for Quadruped, Humanoid, and Hexapod robotics. Integrated YOLOv8 + Whisper + Phi-2 on ARM with sub-50ms ZeroMQ IPC latency. |
|
AI Automation Engineer Architected LLM prompt orchestration pipelines and autonomous agent workflows. Automated 50+ enterprise assets, cutting operational latency by 25%. |
ROBOTICS CLUB @ BIT Bangalore Core organizer for Ideathon 2025 (600+ participants). Leading robotics R&D hardware initiatives, technical workshops, and cross-functional engineering teams. |
Six AI pipelines. One Raspberry Pi 5. Zero Cloud.
Autonomous desktop robot running fully on-device. Real-time vision + speech + local LLM + vector memory simultaneously.
🥉 Best Poster Presentation Award
Sub-200ms LLM query optimization via n-gram pattern mining
Achieved 40–60% token reduction with sub-200ms vector retrieval. Air-gapped architecture with zero telemetry.
Transforming video streams into structured vector memory
Hybrid retrieval pipeline combining PostGIS geospatial indexing with ChromaDB semantic search for AI agent training.
Real-time shopping assistant with custom AI-driven hardware
Deployed vision model running at 20 FPS (92% mAP) on custom ESP32 hardware with 8-hour battery endurance.
🥈 2nd Place @ GlitchVerse 2k25
| Status | Competition / Event | Recognition |
|---|---|---|
| 🥈 | GlitchVerse 2k25 Project Expo | 2nd Place |
| 🥉 | MEMO — Best Poster Presentation | 3rd Place |
| 🏅 | ME-RIISE National Hackathon | 4th Place |
| 🏅 | Anveshana National Engineering Competition | National Finalist |
| 📚 | BIT Bangalore · Robotics & AI | Top 10% · 8.88 CGPA |