Knowledge graph suite - Legal
A suite of tools that helps legal teams understand entity relations in large unstructured corpora
ENGINEERING INTELLIGENT SYSTEMS
I build perception and autonomy systems that work under real-world constraints — combining computer vision, edge ML, and reliable real-time software.
Also building reproducible scientific ML tools and exploring physics-informed neural networks.

A suite of tools that helps legal teams understand entity relations in large unstructured corpora
Reliable visual perception, sensor fusion, and intelligent systems operating on constrained hardware.
Studying latent-space manifolds for solving under-determined family of industrial simulations with single PINN models
Led perception and navigation systems for a drone platform, combining classical computer vision, monocular depth, Deep learning, sensor fusion, and recovery policies for GPS-denied operation.
A configurable experimentation framework for physics-informed neural networks, with reusable problem abstractions, CLI orchestration, and reproducible training workflows.
An AI-assisted document processing project exploring entity extraction, normalization, evaluation, and reviewable structured outputs.
The navigation case study is being documented separately. Repository links above open the corresponding public projects.
Notes on engineering reliable perception pipelines, running reproducible scientific ML experiments, and the trade-offs behind building intelligent systems that work outside a demo.
I'm interested in applied AI, computer vision, robotics, and engineering teams building technically ambitious products.