Yaduraj Singh — Full-stack engineer · AI/ML builder

Ships production-grade systems solo — firmware on ESP32, scalable SaaS, iOS apps. Three live products with real users before 21.

Third-year B.Tech in Computer Science and Engineering, specializing in AI & ML · Graduating 2028

Open to software engineering internships and junior roles. I build full-stack applications, real-time systems and self-hosted infrastructure, with hands-on experience integrating AI into products.

View projects · Download résumé · Contact me

Yaduraj Singh is a 20-year-old full-stack engineer · ai/ml builder based in Dehradun · Greater Noida, India. He designs, builds and operates production software end to end — from firmware running on microcontrollers through to multi-tenant SaaS platforms with real users — and runs the infrastructure underneath it himself.

Highlights

• Self-hosted Linux infrastructure — Nginx, SSL, PM2, systemd. Not just Vercel deploys.
• Manual WebRTC: full ICE/STUN/TURN + signalling layer. No third-party video SDK.
• Designed multi-tenancy, RBAC, caching, CI/CD for a hospital SaaS — solo, from a blank repo.
• Three live production applications with real users before age 21.
• Hackathon-winning hardware + AI system: ESP32-S3 + FastAPI + Flutter, team of two.
• AI/ML pipelines: handwriting synthesis, voice + LLM + TTS cognitive OS, wearable vitals monitoring.

By the numbers

• 2,000+ users in production
• 08 shipped projects
• 03 hackathons won
• 04 self-hosted servers

Now building

Fleet OS (open source · mit) — Git push to the hardware you already own. A Raspberry Pi, an old laptop and a spare VPS, treated as one deploy target — multi-arch builds, constraint-based placement, health-gated rollouts and automatic failover.

• Outbound-only Go agents — no inbound Docker socket, no SSH key, no port to forward. A Pi behind a home router is a first-class node.
• Weighted scheduling across CPU, memory headroom, reliability tier, tags, GPU and affinity — and the placement plan explains every node it rejected.
• Health-gated rollouts: the release that works keeps serving until its replacement proves it can serve too.

Experience

Lead Software Developer — Intern — Gautam Buddha University

2024 · 6 months

• Sole architect of a ground-up institutional ERP — ER diagrams, API contracts, to production deployment.
• Modules: student lifecycle, automated attendance, fee ledger, timetable engine, faculty/admin portals.
• Eliminated 100% paper workflows. ~60% admin overhead reduction.

Stack: Node.js, Express, JWT RBAC, React

Projects

Aarogya Setu

Multi-tenant hospital SaaS — OPD/IPD queues, EMR, appointments, billing, pharmacy, diagnostics. 6-tier RBAC.

Category: SaaS · Self-hosted. Stack: Next.js 14, TypeScript, Prisma, PostgreSQL, Redis, Docker. Live at https://arogya.yaduraj.me.

Tollgate

Cost & usage observability for LLM APIs. Reverse proxy for OpenAI, Anthropic and OpenAI-compatible providers — one base URL for per-feature cost, caching, budgets and runaway-agent alerts.

Category: LLMOps · Proxy. Stack: OpenAI API, Anthropic API, OpenRouter, Vercel. Live at https://tollgate.yaduraj.me/.

MuhDikhai

Omegle-style anonymous video chat. Full WebRTC peer lifecycle built manually over Socket.io. Sub-2s pairing.

Category: Real-time · WebRTC. Stack: Node.js, TypeScript, WebRTC, Socket.io, PostgreSQL. Live at https://batchit.yaduraj.me.

CineVerse

Social film tracking — watchlists, ratings, reviews, discovery feeds. TMDB API across 500k+ titles.

Category: Social · SSR. Stack: Next.js, Firebase, TMDB API, Nginx, PM2. Live at https://cine.yaduraj.me.

SecondMind / CortX

Cognitive OS on ESP32-S3 Sense. Voice → faster-whisper → local LLM → Coqui TTS. VAD firmware, Opus compression.

Category: Hackathon Winner. Stack: ESP32-S3, FastAPI, Flutter, Neo4j, Qdrant. Live at https://cortx.yaduraj.me.

GBU Timetable

Official iOS app for GBU students. Live home-screen widgets via WidgetKit. Real-time schedule updates.

Category: iOS · App Store. Stack: SwiftUI, Combine, WidgetKit. Live at https://apps.apple.com.

Maakosh

Maternal & neonatal health iOS app. Wearable sensor integration (MAX30100, EMG) for real-time vitals. Role-based.

Category: Hackathon Winner. Stack: Swift, Firebase, ThingSpeak.

Bolonyay

Multilingual legal aid iOS app — petitioners with regional advocates. Voice-assisted forms, automated PDF reports.

Category: iOS · Legal Aid. Stack: Swift, Firestore, PDFKit, STT.

Stack

• Languages: TypeScript, Python, Go, Swift, C/C++, Dart
• Web & APIs: Next.js, React, Node.js, Fastify, Express, FastAPI, Tailwind
• Mobile: SwiftUI, UIKit, Combine, WidgetKit, AVFoundation, Flutter
• AI & ML: faster-whisper, Coqui TTS, LM Studio
• Data: PostgreSQL, Redis, Firebase, Prisma, Drizzle ORM, Qdrant, Neo4j
• Infrastructure: Linux, Docker, Docker Buildx, GitHub Actions, Nginx, systemd, PM2, Certbot, Vercel
• Real-time: WebRTC, ICE/STUN/TURN, coturn, Socket.io, Cloudflare Tunnel
• Embedded: ESP32-S3, FreeRTOS, BLE, Opus

Contact

• Email: yadurajsingham@gmail.com
• GitHub: https://github.com/YadurajManu
• LinkedIn: https://www.linkedin.com/in/yadurajenc
• Résumé: https://yaduraj.me/Resume_Web.pdf

For developers and automated agents

A public read-only JSON API exposes everything on this page as structured data. Read the API documentation, the OpenAPI specification, or the llms.txt agent index.

• About — full background and experience
• Developer documentation — endpoints and examples
• Contact — every direct channel
• Privacy — what this site collects