Backend & full-stack developer
I build the parts of a product that have to be right — money maths that reconciles to the cent, APIs that fail loudly, and deploys boring enough to run on a Tuesday night. Django and FastAPI on the server, Postgres underneath, AWS around it.
Open to backend & full-stack roles
↑ liquid glass, 2025
I'm a computer engineering student who wandered onto the server side and never left. The work I like has a correct answer buried in it somewhere — an amortisation schedule that has to balance, a retrieval pipeline that has to cite the document it actually used, a deploy that has to be repeatable.
Most of what I ship is Django or FastAPI behind a small, sharp frontend. I'd rather spend the afternoon on the migration strategy than the hover state. (This page is what happens when I get to do both.)
Currently
neomorphism, 2019Computer
Engineering
Building
DebtClear
Learning
Distributed
Systems
Status
Available
What I write day to day.
Where most of the time goes.
Storage, cache, and the queries between.
Off the laptop, and staying up.
Used where it earns its latency.
Enough to ship the whole thing solo.
Featured build
minimalism, perennialA Django debt-payoff planner. You enter what you owe; it simulates Avalanche and Snowball strategies side by side, charts the payoff curve, and an AI advisor explains the trade-off in the plain language a spreadsheet never will.
2
payoff strategies simulated in parallel — highest-interest-first and smallest-balance-first — so the cost of choosing motivation over maths is visible in rupees, not vibes.
3.3
Llama 3.3 via Groq powers the advisory layer. Groq's inference speed is what makes the advice feel like part of the form rather than a job you wait on.
main on green.Memory and retrieval across Gmail, Calendar, and documents. Vector search and BM25 are fused with Reciprocal Rank Fusion and reranked by Gemini 3 Pro, then a critic sub-agent audits the answer against the chunks it cited — so a claim with no source behind it never reaches you.
TypeScript · Express 5 · MongoDB Atlas · Vertex AI · Cloud Run
Final-year computer engineering student. No employment history yet — what follows is shipped, dated work, most recent first.
| Shipped | Build | What it does |
|---|---|---|
| May 2026 | Mnemos | Production AI memory agent for retrieval across Gmail, Calendar, and documents. Hybrid pipeline merging MongoDB Vector Search and BM25 via Reciprocal Rank Fusion, reranked with Gemini 3 Pro, plus a critic sub-agent that audits every response against its cited chunks to catch hallucinations. Deployed on Cloud Run with SSE streaming. |
| Mar. 2026 | TicketForge | Automation pipeline that ingests raw meeting transcripts and generates structured Linear tickets through a multi-stage Groq prompt chain, extracting tasks, assignees, deadlines, and priorities from unstructured text. Devpost hackathon entry. |
| Feb. 2026 | DebtClear | LLM-powered debt resolution platform with a Negotiate Mode that generates personalised negotiation scripts from a user's debt profile. Django REST on EC2, Dockerised, GitHub Actions CI/CD, Groq (Llama 3) with SSE streaming. Sole developer, production-ready in under 72 hours for Hack the Pulse 2026. |
| Jan. 2026 | DevSync | Full-stack developer portfolio platform — 12+ REST modules, Dockerised with PostgreSQL 16, Redis 7, Celery and Nginx. Local-LLM assistant on Ollama with a ChromaDB RAG pipeline and no external API keys. Jenkins CI/CD with linting, Safety and Trivy scans, and 35+ tests. |
B.Tech, Computer Engineering
Sardar Patel Institute of Technology
Mumbai University · Mumbai, India
Aug. 2023 – June 2027
Computer-science fundamentals with a practical bias: data structures, operating systems, databases, and machine learning, applied to things that actually get deployed.
Featured build
DebtClear
Django · Groq · EC2 · GitHub Actions. Deployed, tested, and live.
4
Builds shipped in 2026
3+
Years on Python
Automated deploys
100%
automated — push to main, tests, ship
1
HTML file, zero requests
Design eras borrowed from
9
one device each, dropped into a maximalist house style
Open to
Backend
roles
Let's build something that has to work.
I'm after backend or full-stack work where the hard part is the system design, the data model, or keeping a deploy boring. If that's what you've got, I'd like to hear about it. Fastest route is email — I reply to everything that isn't a recruiter template.