Building reliable, scalable, and production-ready systems across AI, backend engineering, and real-time applications.
I'm an AI Engineer and Full-Stack Developer passionate about building real-world systems that combine:
- 🤖 Deep learning and LLM integration
- 🏗️ Scalable backend architecture
- ⚡ High-performance real-time applications
- 🔧 Production-grade engineering practices
My work spans LLM evaluation, dataset creation, real-time communication engines, full-stack ML deployment, and open-source tooling. I focus on clarity, performance, and production reliability—shipping software that withstands real-world usage, not just demos.
🎓 Currently: Pursuing B.S. in Data Science at IIT Madras
🚀 Expanding into: Distributed systems, decentralized tech, and high-performance infrastructure
|
|
|
Distributed systems · Concurrency management · Latency reduction (<50ms) · Caching strategies · Performance tuning · Scalable architectures
Engineering Philosophy: Correctness, long-term maintainability, modularity, and real user-tested reliability.
⚡ QuickPoll — Real-Time Voting Engine
High-performance WebSocket system handling 1000+ concurrent users with <50ms latency
- ✅ Strict vote deduplication logic
- ✅ 92% test coverage with comprehensive integration tests
- ✅ Fully automated CI/CD deployments
- �️ Stack: Node.js, WebSockets, Redis, Docker, GitHub Actions
🧠 Brain Tumor Detection — Full-Stack ML Platform
97.94% accurate CNN model deployed as production-ready web application
- ✅ EfficientNet architecture with transfer learning
- ✅ FastAPI backend + React frontend
- ✅ Real-time prediction dashboards
- ✅ Docker containerization + CI/CD pipeline
- 🛠️ Stack: PyTorch, FastAPI, React, Docker, PostgreSQL
🤖 AI-Powered TDS Assistant — Multi-Capability AI System
All-in-one AI assistant for content analysis, automation, and workflow optimization
- ✅ YouTube transcription with Whisper
- ✅ PDF/Excel analysis and data extraction
- ✅ GitHub API integration for automated workflows
- ✅ CLI command execution and public API integration
- 🛠️ Stack: Python, Whisper, LangChain, FastAPI, Gemini API
🎯 Dynamic Prompt Studio — AI Prompt Generation Platform
End-to-end AI tool for intelligent prompt engineering and content generation
- ✅ Gemini API integration for LLM interactions
- ✅ YouTube Data API for content analysis
- ✅ Real-time analytics and metrics tracking
- ✅ Context-aware prompt optimization
- 🛠️ Stack: Python, Gemini API, FastAPI, React, WebSockets
� Portfolio LLM Chatbot — Fine-Tuned Conversational AI
Custom-trained Phi-2 model for portfolio project retrieval and interaction
- ✅ Fine-tuned on 40+ project descriptions
- ✅ Context-aware retrieval with RAG pipeline
- ✅ Custom React UI with real-time responses
- ✅ Semantic search and intelligent filtering
- 🛠️ Stack: Phi-2, LangChain, Transformers, FastAPI, React
🔗 Snapbit — First Open-Source Contribution
Significant improvements to mobile experience, authentication, and UI consistency
Contributions:
- 📱 Enhanced mobile responsive design and touch interactions
- 🔐 Improved authentication workflows and session management
- ⚡ Optimized state handling and component performance
- 🎨 Strengthened UI consistency across the platform
Impact: This contribution deepened my expertise in collaborative engineering, code review quality, and production-oriented frontend optimization.
- 🤖 AI/ML tooling and developer utilities
- 🏗️ Distributed systems and infrastructure projects
- 🔧 Stability-focused backend frameworks
- 🌐 Decentralized application components
Currently exploring:
- ⛓️ Blockchain fundamentals and consensus mechanisms
- 🏗️ Distributed architectures and protocol design
- 🔮 Oracle patterns and off-chain computation
- 🛡️ Protocol stability and security best practices
Leveraging existing strengths:
My foundation in APIs, concurrency management, containerized deployment, and real-time systems positions me well for building decentralized applications and off-chain services.
timeline
title Career & Education Journey
2024 : Started B.S. Data Science (IIT Madras)
: Built QuickPoll Real-Time Voting Engine
: Deployed Brain Tumor Detection ML Platform
: Created HTML Citation Cleaner Tool
2025 : AI-Powered TDS Assistant (Multi-capability AI)
: Dynamic Prompt Studio (LLM Integration)
: Portfolio LLM Chatbot (Fine-tuned Phi-2)
: Mercor.io · Soul AI · Outlier (LLM Evaluation & Datasets)
: First Open-Source Contribution (Snapbit)
: Expanding into Decentralized Systems
| Year | Achievement | Impact |
|---|---|---|
| 2024 | 🎓 Enrolled at IIT Madras | B.S. Data Science program |
| 2024 | ⚡ QuickPoll Launch | 1000+ concurrent users, <50ms latency |
| 2024 | 🧠 Brain Tumor Detection | 97.94% accuracy, production deployment |
| 2025 | 🤖 AI Assistant Suite | 3 major AI-powered tools shipped |
| 2025 | 💼 Industry Experience | LLM evaluation, dataset creation, SFT |
| 2025 | 🌱 Open-Source Start | Contributing to production codebases |
|
ML/DL · LLM · RAG |
Backend · Frontend · DevOps |
Complete Profile |
Every system should be reliable, understandable, and ready for scale.
┌─────────────────────────────────────────────────────────────┐
│ Correctness → Maintainability → Scalability → Impact │
└─────────────────────────────────────────────────────────────┘