I'm an AI Robotics practitioner. My day-to-day work revolves around expanding my knowledge in Embedded Systems. I'm open to collaborate on interesting projects.
| Project | Description | Technologies Used / Learned |
|---|---|---|
| Greenhouse Segmentation (GIS + ML) | A reproducible and replicable ML pipeline for desired object detection by fine-tuning segmentation model on geospatial imagery end to end Georeferenced for GIS tasks like reverse geocoding(eg. aadresses of the detected objects | PyTorch, DuckDB, RasterIO, GeoPandas, leafmap, QGIS/ArcGIS, GCP, Data processing, Architectures and Encoders likes UNet++, EfficientNet |
| RAS-ROS2 | Robot Application Stack for ROS 2 developed as part of my role at Open Science Stack. Focused on modular robotics software, integrated with simulation and hardware. | xArm (UFactory), Franka Emika, KUKA, ROS 2, MoveIt 2, Docker, C++, Python, Gazebo, Git, CI/CD |
| Behavior Tree ROS 2 Middleware (RAS-ROS2) | Custom middleware layer built on Behavior Trees using BehaviorTree.CPP library to coordinate perception, planning, and control nodes, improving modularity and runtime adaptability in RAS-ROS2. | xArm (UFactory), ROS 2, BehaviorTree.CPP, C++, Python, RViz |
| Aruco based Grid System (Tic Tac Toe) | AI-driven vision module using ArUco markers to detect and map a Tic Tac Toe board, enabling the robot to perceive moves, plan responses, and play in real time. | xArm (UFactory), Franka Emika, OpenCV, ArUco, Python, ROS 2, CUDA, PyTorch, 3D printing and modeling |
| Open-Sourcing RAS-ROS2 | Released the RAS-ROS2 stack to the public during ROSCon 2024, making it freely available for the global robotics community. | GitHub, ROS 2, Open Source Best Practices |
| Drone Systems at IISc | Built and tested drone systems for research applications during my internship at the Indian Institute of Science. | PX4, MAVROS, Nutty RTOS C++, Python |
| ARC-Net Research Fellowship | Explored Embedded Systems to applied robotics research under the ARC-Net program, a collaborative initiative between IISc and ARTPARK. | Franka Emika, Kawada NEXTAGE, Embedded C, Nutty RTOS, ARM Cortex-M, STM32 |
| Live Footfall system | Developed a live footfall counting system for event venues, optimized for real-time inference on edge devices. | TensorRT, PyTorch, OpenCV, YOLOv8 |
| Project | Description | Technologies Used |
|---|---|---|
| Elective Selection Web App (Alliance University) | Internal Modern cloud based web app for course elective selection. Check it out | Flutter, Firebase |
| Hasti Ganga Agro Industries Website | Static site generator (SSG) for Hasti Ganga Agro Industries company. Check it out | Astro, Cloudflare |
| Skin6 Clinic Website | WordPress website for Skin6 clinic. Check it out | WordPress, Hosting, SEO |
| Organization | Role | Duration |
|---|---|---|
| Audria | Hardware Lead | September-2025 to Present |
| Indian Institute of Science (IISc) | Drone Intern | March-2024 to August-2024 |
| ARTPARK (AI & Robotics Technology Park) | Robotics Intern | August-2024 to January-2025 |
| Open Science Stack | Robotics Engineer (RAS-ROS2) | September-2024 to December-2024 |
| ARC-Net (IISc + ARTPARK) | Research Fellow | January-2025 to July-2025 |
| Alliance University COE CV Lab | Computer Vision Intern | Part time |
Robots
| KUKA | Kawada NEXTAGE | xArm (UFactory) | Franka Emika |
|---|
Programming Languages
 |
 |
 |
 |
 |
|---|---|---|---|---|
| Python | JavaScript | Assembly | C | C++ |
Libraries
 |
 |
 |
 |
 |
 |
|---|---|---|---|---|---|
| TensorFlow | Docker | Terraform | PyTorch | OpenCV | CUDA |
Frameworks
 |
 |
 |
 |
 |
 |
|---|---|---|---|---|---|
| Flask | ROS 2 | MoveIt | React | Astro | Flutter |
Infrastructure
 |
 |
 |
 |
 |
 |
 |
|---|---|---|---|---|---|---|
| GCP | Kubernetes | AWS | Cloudflare | Firebase | RabbitMQ | Redis |
Databases
 |
 |
|---|---|
| MongoDB | DuckDB |
Tools
 |
 |
 |
 |
 |
|---|---|---|---|---|
| Ubuntu | ZSH | VS Code | Git | Jupyter Notebook |
Hardware / Architecture
 |
|
|---|---|
| ARM | ARM Cortex-M |
Embedded & Wireless
 |
 |
 |
|
|
|
|---|---|---|---|---|---|
| nRF5340 | Zephyr RTOS | BLE | nRF Connect SDK | BLE Audio (LC3) | DMIC / I2S |
Segmentation Models
| Unet | Unet++ | FPN | PSPNet |
|---|---|---|---|
| DeepLabV3 | DeepLabV3+ | Linknet | MAnet |
| PAN | UPerNet | Segformer | DPT |
Encoders (Backbones)
| Mix Vision Transformer (MiT) | MobileOne | MobileNet | EfficientNet |
|---|---|---|---|
| ResNet | ResNeXt | SENet | DPN |
| VGG | DenseNet | Xception | Inception |
Optimization & Evaluation
| AMP (mixed precision) | Cosine LR | Dice + BCE losses |
|---|---|---|
| IoU / F1 / Precision / Recall | Early-stopping | Checkpointing |
 |
 |
|---|