Building intelligent systems at the boundary of language model alignment, retrieval architecture, and high-performance computation. Scroll to traverse the landscape — or hit Pilot Mode to fly freely through the world.
Specializing in high-performance AI systems where complex algorithms meet efficient machine execution.
Most COPD screening tools rely on a single modality. This system fuses lung sound recordings, clinical survey data, and chest imaging into a unified deep learning pipeline. Lightweight, built on public clinical datasets — designed for reliable early detection where resources are constrained.
Transforms unstructured PDFs into interactive knowledge bases. FAISS vector stores with all-MiniLM-L6-v2 embeddings power high-precision semantic retrieval. Citation-backed answers with page-level metadata eliminate hallucinations.
Watches code evolve through pull requests and injects analytical signals — Big-O complexity detection, vulnerability scanning, style enforcement. Integrated with GitHub REST API to trigger automated PR feedback.
High-velocity code analysis — scanning in real-time, identifying structural inefficiencies, emitting refactoring signals via REST. Hot-path profiling and efficient STL data structures drive pattern matching at machine speed.
Open to AI/ML roles, research collaborations, Computer Vision Engineer, Backend Architect and projects that demand systems thinking. Every message receives a thoughtful response.
Signal received.
I'll respond to within a day.