SIDDHARTH.EXE
Hey — I'm Siddharth. Scroll down and I'll narrate, or ask me anything below.
I'm an AI & Machine Learning engineering student at BNM Institute of Technology, and I build practical AI systems end-to-end. My work spans LLM fine-tuning dataset forensics, explainable space mission-assurance, AI-powered candidate verification, medical imaging with Vision Transformers, offline academic intelligence, anonymous mental-wellness platforms, edge-AI rider safety, payment data quality, local voice AI, privacy tools, and blockchain-based environmental impact tracking.
I like taking a hard research problem all the way to something real people can use — combining models, automation, full-stack development, local LLMs, and infrastructure into products that are practical, scalable, and meaningful.
An offline-first Academic Intelligence System with hybrid retrieval, local Ollama generation, and a full study workspace — no cloud, no subscriptions.
NirmiqResearchOS is an offline-first, privacy-conscious academic intelligence system. It features a FastAPI backend with hybrid BM25 + Chroma vector retrieval (RRF reranking), and a custom Next.js study workspace with 10 study modes — from deep research to exam answers and revision notes. Documents (PDF, text, markdown, images) are ingested locally, chunked, and indexed with page-level caching. Generation and embeddings run through Ollama. Zero cloud dependency.
A selection of AI, automation, research, privacy, health, and blockchain systems I've built or am actively building.
Hackathon results and certificates from the past few months — built and shipped within 24–48 hour windows.
Ways I can support AI, automation, and product-building projects.
Open to internships, AI/ML projects, collaborations, research opportunities, freelance automation work, and product-building opportunities.