$ whoami

Rishik Garg

Backend Engineer

Building software that survives beyond the demo.

Currently debugging:
Life, distributed systems, and side projects that got way bigger than planned.

$ git log --author="Rishik"

feat: build another backend
fix: forgot semicolon
docs: write 800 lines of documentation
refactor: because it bothered me

"If miracles only happen once, what are they called the second time?"
— Kisuke Urahara, Bleach

I'm a Backend Engineer focused on building secure, scalable, and maintainable applications.

Most of my work revolves around Spring Boot, while FastAPI is my go-to whenever AI or data-intensive workflows are involved.

I enjoy building complete systems—from backend architecture and APIs to deployment and documentation—because good software should still be understandable long after it's shipped.

A productivity platform built to make consistency easier, not productivity noisier.

Spring Boot • React • PostgreSQL • Spring Security • JWT • OAuth2 • WebSockets • Docker

LifeOS began as a task manager and gradually evolved into a full productivity ecosystem featuring secure authentication, real-time communication, analytics, AI-assisted task management, and comprehensive documentation.

Highlights

• 🔐 JWT Authentication & Google OAuth
• ⚡ Real-time notifications with WebSockets
• 📊 Activity tracking, streaks, levels & leaderboards
• 🤖 AI-assisted task generation
• 🐳 Dockerized deployment
• 📚 Architecture-focused documentation

Exploring how data and AI can help cities make better traffic management decisions.

FastAPI • React • Llama 3 • Groq • H3 • Pandas

Developed during Flipkart Gridlock 2.0, RecLog analyzes parking-induced congestion, identifies traffic hotspots using geospatial analytics, and recommends resource allocation through explainable AI. The project introduces the Priority Congestion Risk Index (PCRI) to prioritize intervention.

Highlights

• 🗺️ H3-based geospatial clustering
• 🧠 Priority Congestion Risk Index (PCRI)
• 🤖 Explainable AI recommendations
• 📈 Resource allocation simulator
• 📊 Interactive analytics dashboard

Making handwritten prescriptions a little easier to understand.

FastAPI • Google Vision • Hugging Face • Gemini

Medify combines OCR, biomedical NLP, and AI to extract prescription information, verify medications, analyze drug interactions, and generate patient-friendly summaries through a multi-stage processing pipeline.

Highlights

• 📄 OCR-powered prescription extraction
• 🧬 Biomedical Named Entity Recognition
• 💊 Drug interaction & dosage verification
• 📝 AI-generated patient summaries

I enjoy building software that's designed to outlive the initial demo.

Over time, I've found myself caring less about adding more features and more about building systems that are easier to understand, extend, and maintain.

Some principles I try to follow:

• 🧩 Design for maintainability before complexity.
• 🔐 Treat security as a feature, not an afterthought.
• 📖 Write documentation that future contributors (and future me) can actually use.
• 🤖 Use AI where it genuinely improves the product.
• 🚀 Build complete applications—from architecture to deployment.

"Code gets you to version 1. Good engineering gets you to version 10."

Currently exploring and building around:

• 🏛️ Backend architecture & system design
• ⚡ Performance optimization and caching
• 📡 Distributed systems
• 📊 Observability & monitoring
• 🐳 Production-ready deployment workflows

Idea
  ↓
Architecture
  ↓
Backend
  ↓
Testing
  ↓
Frontend
  ↓
Deployment
  ↓
Documentation
  ↓
Refactor

When I'm not building projects, you'll usually find me:

• Reading about backend architecture and distributed systems
• Improving existing projects instead of starting new ones
• Exploring tools and workflows that make software development more enjoyable

If you'd like to talk about backend engineering, system design, AI-assisted applications, or collaborate on interesting projects, I'd love to connect.

📧 Email: rishikgarg.dev@gmail.com

💼 LinkedIn: https://linkedin.com/in/rishik-garg-dev

"I enjoy building software that people can rely on—and code that future developers won't hate."

Thanks for stopping by! 👋