Work

Things I've designed and built.

A collection of products, interfaces, and experiments. Each card loops a short demo — hover to explore, or open a case study for the full story.

Bengaluru Vega demo
2nd Prize · ₹1.5 Lakh (≈ $1,700)
2025Machine Learning

Bengaluru Vega

Predicting multi-modal ETAs.

A machine learning system that predicts accurate arrival times across buses, metro, and road travel for Bengaluru's transit network — fusing multiple data sources into one reliable estimate.

PythonScikit-learnPandasML
View case study
PySub demo
2nd Prize · ₹25,000 (≈ $300)
2025Robotics · Controls

PySub

Control for a 6-DoF ROV.

A Python control stack for an 8-thruster V-config underwater ROV, implementing PID-based stability, thrust allocation, and live telemetry for precise 6 degrees-of-freedom maneuvering.

PythonPID ControlRobotics
View case study
Autonomous Surface Vehicle demo
2026Controls · Networking

Autonomous Surface Vehicle

Self-navigating on water.

End-to-end control software for an autonomous surface vehicle — covering networking, control loops, and carrot-chasing path-following algorithms for reliable autonomous navigation.

PythonGuidanceNetworking
View case study
Hyperspectral Pipeline demo
2025Software · Research

Hyperspectral Pipeline

Compression for satellite payloads.

A lightweight, memory-efficient pipeline for compressing, encoding, and reconstructing hyperspectral image (HSI) data for satellite payloads, with automated batch processing to accelerate research workflows.

PythonNumPyImaging
View case study
16-Bit Basic Computer demo
2024Computer Architecture

16-Bit Basic Computer

A CPU, from logic up.

A 16-bit basic computer (Mano Machine) built from digital logic — registers, buses, and a control unit — to explore computer architecture from first principles.

Digital LogicArchitectureVerilog
View case study
Visual Odometry Pipeline demo
2024Computer Vision

Visual Odometry Pipeline

Real-time trajectory tracking.

Developed a robust visual odometry system for autonomous navigation using stereo camera setups. The pipeline includes feature extraction, matching, outlier rejection, and camera pose estimation in GPS-denied environments.

C++OpenCVROS
View case study