🚁
Vision-Language Navigation
Real-time AI on Edge Devices
Multi-model perception system that runs entirely on low-cost hardware. Zero cloud dependency — all AI processing happens on-device.
I build AI solutions that actually work in production. Whether it's computer vision, NLP, or cutting-edge LLMs — I take ideas from concept to deployment, solving real business problems along the way.
Bridging the gap between cutting-edge research and real-world deployment
I help businesses turn complex AI ideas into working products. From computer vision to natural language processing, I specialize in building AI systems that actually work in production — not just in papers. My approach: prototype fast, iterate quickly, deliver results.
From research to deployment — a full-stack AI toolkit
Research papers, real-world deployments, and cutting-edge innovations
Real-time AI on Edge Devices
Multi-model perception system that runs entirely on low-cost hardware. Zero cloud dependency — all AI processing happens on-device.
Intelligent Drone Vision System
An interactive "flying eye" capable of perceiving, understanding, and responding using vision and language models for autonomous surveillance.
Advanced Identity Verification
Deepfake detection with liveness verification and explainable AI for secure real-time authentication.
Multi-Sensor Assistive Navigation
Published research and a granted patent on AI-integrated multi-sensor assistive navigation with multimodal feedback.
Sim2Real Transfer Pipeline
Autonomous navigation in ROS with sensor fusion (LiDAR, camera, IMU), occupancy grid mapping, and adaptive pathfinding (A*, RRT*).
Contributing to the advancement of AI and robotics research
IEEE CASE 2026 • First author
Anomaly detection running onboard a UAV, with no examples of the anomaly class used either in training or in fitting the decision threshold. Validated in flight against a geographically disjoint benchmark.
MMLoSo 2025, ACL Workshop • pp. 86–94
Enhanced accessibility using lightweight VLMs with custom evaluation frameworks for on-device deployment.
For Visually Impaired Individuals — Multi-sensor fusion with real-time health monitoring and AI-driven navigation.
IEEE RA-L, in preparation • First author
A 0.17M-parameter geometry expert that lands within 2.0 percentage points of a fine-tuned 111.6M-parameter vision-language model, running in real time inside a 10W budget.
ICANN 2026, Springer LNCS • Co-author
International wins (and proof I can also lead people outside of terminals)
International Conference on Social Robotics
Odense, DenmarkWinner
IIT MandiAdvanced UAV Systems Training
IIT MandiRaj. Dte. SNIC CAMP
Jaisalmer, 2022College Team Leadership
ECBAnimal Welfare Society Volunteer
Active MemberOpen to research collaborations, AI/ML consulting, internships, and full-time opportunities. Let's discuss how I can help with your next project.