01 — Experience & selected builds
A few places where I’ve turned ambiguous problems into shipped systems.
Built a computer vision pipeline for detecting and segmenting defects in dental X-rays, merging four open-source datasets into one standardized training and evaluation flow.
Built the AI foundations for an education platform: structured generation, subject-specific reasoning pipelines, and RAG systems that route work across models.
Built the Python and SQLite backend for the Indian Fertility Society website, including structured data, admin workflows, and validated APIs for dynamic content.
02 — Side quests
From a pocket-sized assistive device to infrastructure that stays on its feet.
An assistive system that reads handwritten text aloud and describes surroundings — running end-to-end on a single Raspberry Pi 5.
A Go load balancer in front of a Python backend with request validation, leaky-bucket rate limiting, and cached game state.