Siddharth Surange
Retrieval that cites its sources instead of hallucinating them. Agents that fail loudly instead of silently. 10 years in software, the last 2 shipping LLM-powered products to production — not just notebooks.
Automated AI research briefing agent — monitors Google AI, DeepMind, OpenAI, Anthropic, arXiv and more, then delivers a curated daily digest to Telegram with follow-up Q&A over a 14-day rolling knowledge base.
Dual-layer dedup (SHA-256 + cosine), tiered source ranking, cited RAG answers — ~$8/month in production.
Scans official and top community Claude Code repos for agents, skills, commands, and hooks, then lets you interactively install what you want straight into ~/.claude/ — with license tracking and attribution built in.
8 curated sources with per-repo license classification and a unified scanner across heterogeneous repo layouts.
A team-ops layer for Cursor — versioned rules, agent skills, and git guardrails that roll out across an entire engineering team in minutes, not sprints.
16 versioned skills, 7 enforcement rule packs, 3 automated guardrail hooks, team policy baked in.
HR-focused resume analysis that runs entirely on local LLMs — no API keys. Extracts structured data from PDFs, flags missing sections, and generates tailored interview questions.
Fully offline with 8-bit quantized Llama 3.1/3.2 and IBM Granite.
Monorepo of production-grade experiments: a Chrome extension + FastAPI/Qdrant backend for semantic search over your browsing history, and a deployed agentic app with web search and image generation.
Shipped as working demos, not notebooks.
I write on Medium about the parts of GenAI that are underdocumented.
Locally first — LM Studio and quantized Llama/Granite models — with cloud APIs only once an idea is proven. Right now I'm deep in agentic dev tooling (Claude Code, Cursor), RAG that's actually evaluated (dedup, ranking, citations, before/after evals), and local quantized LLM setups.
Certified: Oracle Generative AI Professional · Google Cloud Professional Data Engineer. Experiments ship to Hugging Face Spaces; the write-ups land on Medium.