Professional Experience

Researching event-based star tracking for high-speed spacecraft attitude estimation under low-SWaP constraints, with emphasis on centroid correction, adaptive uncertainty, and evaluation-driven state estimation.

• Lead research on the ASU-Alphacore NASA event-based star-tracking project for high-speed spacecraft attitude estimation under low-SWaP constraints.
• Led development of a centroid-correction algorithm, improving the measurement model from brightness-only correction to a speed-aware and brightness-aware formulation for fast-moving stars.
• Implemented speed-aware centroid correction with adaptive measurement uncertainty, event batching, centroid extraction, astrometric initialization, and weighted Wahba attitude estimation.
• Validated the prototype on synthetic star-motion tests and public real-night-sky resources; compared baseline and speed-aware runs on a 3.0 s sequence with 372 batches and about 8.7M events while preserving runtime feasibility.
• Supported the Phase I proposal submitted to NASA and currently contributing to Phase II technical progress, including experiments, evaluation design, and research/proposal writing.

Built backend and data-intensive software for large, distributed defense platforms with strict reliability, availability, and performance requirements, including Integrated Electronic Warfare Systems and naval projects.

• Led backend development for the Indian Army's Integrated Electronic Warfare Systems (IEWS), delivering distributed C/C++ and Python services, data pipelines, and control-plane components across heterogeneous hardware; managed a team of 6 engineers.
• Designed and documented a reusable Health Monitor SDK for Indian Navy programs (3D-ASR, CMS), standardizing telemetry, diagnostics, fault-detection interfaces, and observability across products.
• Automated Fire Controller Units for Dhanush, ATAGS, and Pinaka platforms, reducing manual intervention by 75% through deterministic control logic, Python scripting, and repeatable validation workflows.
• Implemented high-performance data cataloging and processing features in the Ballistic Logix library for ballistic range data, improving processing efficiency by 60% and enabling downstream analytics and model development.

Built machine-learning and computer-vision-driven automotive interfaces with real-time data exchange and visualization tooling.

• Developed a gesture-driven car infotainment prototype using OpenCV and TensorFlow, integrating REST APIs for real-time data exchange and achieving 92% gesture-recognition accuracy.
• Combined machine learning, computer vision, and data visualization for a real-time automotive interface; profiled and debugged the perception pipeline to resolve latency and performance bottlenecks.