const manmay = {
  focus:    "backend systems, real-time infra, AI integration",
  degree:   "B.Tech CSE (AI & ML) — VIT Chennai, 2028",
  now:      "building NexCare, a telemedicine platform with video + pharmacy workflows",
  looking:  "backend / AI internships — available from May 2026",
  belief:   "a well-designed API is better than a pretty UI",
};

I'm a backend-heavy full-stack developer who prefers designing systems over cloning tutorials. Most of my time goes into thinking about how data flows, where things break under load, and how to make services talk to each other without a mess. I work primarily in Node.js and Go on the backend, use React/Next.js when the frontend actually needs me, and reach for LangChain + FastAPI when there's an AI problem worth solving.

Things I built to solve actual problems, not to pad a portfolio.

Go WebRTC WebSockets PostgreSQL Redis Docker

Full-stack telemedicine system with real-time video consultations, asynchronous prescription management, and a pharmacy workflow that tracks medication status from prescription to dispensing.

The interesting engineering problems: WebRTC signalling under unreliable network conditions, session state across disconnects, and keeping the pharmacy queue consistent without a race condition when multiple pharmacists are online.

Status: Active development  |  Repo →

FastAPI LangChain React MongoDB Plotly Recharts

Built in 24 hours at a hackathon. Agentic AI system that takes a disease query, hits biomedical databases (PubMed, ChEMBL), and surfaces drug-repurposing candidates with evidence summaries. LangChain agents handle query decomposition and source aggregation; Plotly renders the compound relationship graphs.

This was the project that made me take LLM tool-use seriously.

Status: Hackathon build  |  Repo →

MongoDB Express.js React Node.js Cohere AI

Quiz platform where you create a test and share a unique ID — anyone with the link takes the same test instance. Cohere AI generates questions from a topic + difficulty input. Built proper analytics (per-user scores, question-level accuracy, time-per-question) rather than just showing a final number.

The schema design for tracking attempts without duplicating question data was the most interesting part.

Status: Complete  |  Repo →

Node.js Express.js MongoDB Mapbox API Passport.js

Yes, it's an Airbnb-style app. But the point wasn't the idea — it was getting comfortable with geospatial queries, RESTful resource design, and authentication flows (session vs token, CSRF handling). Mapbox integration for radius-based listing search was the non-tutorial part.

Status: Complete  |  Repo →

I use what solves the problem. These are the tools I'm actually comfortable with:

Backend   Node.js Express.js Go (Gin) FastAPI WebSockets REST gRPC (learning)

Frontend   React Next.js TypeScript Tailwind CSS

Data   PostgreSQL MongoDB Redis Supabase MySQL

Infrastructure   Docker CI/CD (GitHub Actions) Linux

AI / ML   LangChain Cohere AI FastAPI Python

Languages   JavaScript TypeScript Go Python C++ Java

• 🏗️ Building NexCare — telemedicine + pharmacy workflow platform
• 📖 Going deeper into distributed systems — consensus, replication, partition tolerance
• ⚙️ Learning gRPC for inter-service communication in Go
• 🌍 Actively applying for backend / AI engineering internships — available from May 2026

Software Engineer Intern — Cestrum Technologies (Nov 2025 – Feb 2026) Built WebSocket backend for real-time communication — connection lifecycle management, message broadcasting across concurrent sessions, and server-side event processing. First time I had to care about what happens at 500 simultaneous connections, not just 5.

Web Development Lead — CodeChef VIT Chennai (Jul 2025 – Present) Led backend for SIH 2025 team project in Go + Next.js + MongoDB. Mentored juniors on API design and database schema decisions.

Technical Team — E-Cell VIT Chennai (Aug 2025 – Apr 2026) Built a real-time Stock Market Simulator — live price indicators, interactive trading UI, event-state synchronization.

Set up WakaTime + the waka-readme GitHub Action to auto-populate live coding stats here.

If you're building something interesting in backend infrastructure, real-time systems, or AI tooling — and you need someone who'll figure things out rather than wait to be told what to do — reach out.

imanmay2@gmail.com  |  LinkedIn  |  Available from May 2026