B.Tech CSE (AI & Data Science) · Nxtwave Institute, Pune · India 🇮🇳
harshit = {
"role" : "AI Safety Engineer | Applied AI Systems",
"focus" : ["Agentic AI", "Multi-agent Orchestration", "AI Evaluation", "On-device ML"],
"building" : ["AIOS"],
"researching": ["PTF — Agentic AI Failure Study (SSRN)", "Genesis Lock — Blockchain AI Governance"],
"philosophy" : "From math → production. No API wrappers, no shortcuts.",
"location" : "India 🇮🇳 — India-first deployment always.",
}|
Empirical study of an agentic AI failure incident An investigation into a real multi-agent system failure — root cause analysis, failure cascade mapping, and safety mitigations. Submitted to SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6472278
|
Blockchain governance for self-evolving AI A governance proposal for constraining and auditing self-modifying AI systems using on-chain commit structures and quorum-based approval gates.
|
| Project | What | Status |
|---|
| 🖥️ AIOS | Local offline AI desktop assistant + AI-Native Operating Environment — zero cloud dependency | ✅ Live |
| Project | Signal | |
|---|---|---|
| ⚡ | Anthropic Performance Challenge | 10.2× kernel speedup on Anthropic's real hiring take-home. VLIW cycles: 147K → 14K |
| 🧠 | micrograd-JAX | Rebuilt Karpathy's autograd in JAX — grad, jit, vmap, XLA-compiled, GPU-ready |
| 🖥️ | AIOS | Local AI desktop assistant. Ollama-powered. Built before on-device AI was a trend |
| 📚 | ML Fundamentals | 19 notebooks — every core ML algorithm from scratch in NumPy. #100DaysOfML |
| 🏭 | CodeAlpha ML Internship | Production ML: credit scoring, image recognition, disease prediction, speech emotion |
Languages
AI / ML
Tools & Infra
[ ] PTF Paper — peer-reviewed ArXiv publication
[✓] PTF PAPER(Preprint) - Published to SSRN
[✓] Open source: 10+ merged PRs across AI safety / NLP repos
[✓] AIOS — shipped
[✓] Anthropic Performance Challenge — 10.2× speedup
[✓] micrograd-JAX — XLA-compiled, GPU-ready
[ ] Building a small prototype to the PTF and genesis lock frameworks and around "How humans can control AI "
I'm actively looking for ML Engineer / Applied AI / AI Safety / AI Systems roles — remote globally or India-based.
If you're building in agentic AI, multilingual NLP, on-device inference, or AI safety — let's talk.
"Build from first principles. Ship to production. Leave the APIs for later."