Available for Full-Stack roles

Ahan
Hegde

Full-Stack & ML Engineer with 4 years building production systems — from real-time CV pipelines to LLM-powered backends and semantic search engines.

See my work Get in touch
Background

The human behind the commits

I'm a 25-year-old Full-Stack & ML Engineer based in Bangalore. Over four years I've shipped real features at a London-based SaaS startup (Flixmedia), a semiconductor AI company (Rayvector / Lam Research), and an early-stage healthtech startup — always across the full stack.

I gravitate toward problems that sit at the seam between ML and systems.

Certified in Deep Learning (Andrew Ng / Stanford) and Mathematics for ML (Imperial College London).

Core Stack

Python FastAPI JavaScript TypeScript React Node.js PHP

ML / AI

PyTorch TensorFlow OpenCV FAISS Transformers / BERT LLM APIs

Infra & Tooling

Docker GitHub Actions BSD Sockets REST APIs PostgreSQL CI/CD
Career

Where I've worked

Rayvector Tech
Bangalore
ML & Backend Engineer
Built AI systems for semiconductor manufacturing at Lam Research USA. Shipped a RAG-based document bot (Mistral + FAISS), an OCR pipeline for serial number capture, a real-time camera focusing interface using OpenCV, and a BSD socket communication module for air-gapped IoT devices.
Flixmedia / Corvid
London (Remote)
ML Engineer
One of 5 engineers on the AI team at Flixmedia. Built a video localisation pipeline (Whisper + M2M100 + Google TTS), an AI image curation service used by Samsung, LG, Intel, and a 10× faster metadata extraction pipeline. Full CI/CD with GitHub Actions.
ekadyo.in
India
Lead Full Stack Developer
Led full-stack development of a medical chatbot platform. Built symptom-checking chatbots (500 diseases), a PubMed semantic search engine over 1M+ articles using BERT embeddings, a diabetes risk chatbot for a hospital chain, and a pharma concept browser — all end-to-end, from UI to API to model integration.
Selected Work

Things I've built

💬
NormieChat
A full-stack chat application built from scratch. Live, deployed, and actually used by people.
Full-Stack Real-time
↗ normiechat.in
🔬
RAG Bot for Semiconductor Docs
Custom document Q&A system. Mistral-powered response engine, FAISS vector indexing, FastAPI backend — built to actually work in a semiconductor fab environment.
RAG Mistral FAISS FastAPI
🎥
Video Localisation Pipeline
End-to-end pipeline to auto-localise product videos for European retailers. Stitched together Whisper, Meta's M2M100, and Google TTS into a clean backend microservice.
Whisper M2M100 TTS Python
📷
Real-time Camera Focus Interface
Live video analysis tool for adjusting TIS Camera focus rings during silicon wafer monitoring. OpenCV at the core, performance-critical sections rewritten in C++ to cut latency.
OpenCV C++ Real-time
🧬
PubMed Semantic Search Engine
Deep semantic search across 1M+ PubMed articles using BERT-based embeddings. Full stack: search UI, results interface, filtering, and the backend APIs powering it all.
BERT Semantic Search Full-Stack
🖼️
AI Image Gallery Curation
Backend service that auto-selected best product images for brands like Intel, Samsung, LG, and Sony by removing dupes and ranking assets via similarity detection.
Computer Vision Microservice Python
Contact

Let's talk

Open to full-stack engineering roles. I work best on teams that care about what's actually being built, not just the stack it's built on.

ahansnash@gmail.com +91 63612 11545 GitHub LeetCode