I'm an Electronics and Communication Engineering undergraduate at IIT Patna who enjoys turning difficult, ambiguous problems into systems people can actually use. My recent work spans safety-first desktop agents, edge speech models, geospatial decision intelligence, local-first storage, and interactive 3D experiences.
I care about three things in particular: measurable engineering, clear product thinking, and responsible automation.
LinkedIn · Email · Cinematic portfolio
| Project | What it solves | Engineering focus |
|---|---|---|
| Meridian | Converts geopolitical energy shocks into ranked, compliance-checked procurement responses | Next.js, TypeScript, deck.gl, transparent simulation, multi-agent reasoning |
| tiny-kws | Recognizes spoken commands on constrained hardware with a 119K-parameter model | PyTorch, audio ML, reproducible evaluation, 96.65% test accuracy |
| ARK | Builds a searchable, deduplicated backup vault without modifying source files | Python, SQLite/FTS5, content-addressed storage, safety invariants |
| BEAM | Transfers files privately across devices over the local network | Rust, TLS, LocalSend v2 interoperability, Swift and Android bindings |
| SENTINEL | Automates macOS workflows under strict model-budget and safety constraints | TypeScript, macOS Accessibility, local memory, guarded execution |
| APEX | Turns opportunities into truthful, tailored, review-first application workflows | Next.js, Prisma, LaTeX, browser extension, constrained LLM pipelines |
At Windflow.ai, I worked as an AI Engineering Intern on production and evaluation infrastructure:
- shipped an LLM voice interviewer on GCP Cloud Run;
- unified 14 OCR engines and 9 LLM extractors behind a 250+ test benchmark suite;
- evaluated a 20-model forecasting zoo across 12 public datasets;
- built human-in-the-loop aggregation using ROVER voting, Dawid–Skene EM, and LLM arbitration;
- instrumented per-call LLM cost analytics with Langfuse, LiteLLM, and PostgreSQL.
Languages Python · TypeScript · JavaScript · C++ · C · Rust · SQL · Bash
AI / ML PyTorch · LLM pipelines · agentic systems · speech · forecasting · evaluation
Applications Next.js · React · Node.js · FastAPI · PyQt5 · Prisma
Data PostgreSQL · SQLite · content-addressed storage · geospatial systems
Infrastructure GCP · Docker · GitHub Actions · Langfuse · Linux/macOS tooling
- Selected for Amazon ML Summer School 2026 — top 3,000 from 134,000+ applicants.
- Bronze medalist at the Inter IIT Cultural Meet 7.0 in Stageplay.
- TEDxIITPatna planning and curation sub-coordinator; selected 8 speakers from 1,000+ prospects.
- Interested in AI systems that are efficient, inspectable, and grounded in real user constraints.
I'm open to conversations around AI engineering, applied ML, agent infrastructure, developer tools, and ambitious product prototypes. The best way to reach me is LinkedIn or email.