Rahul Balaji

Real-time AI-powered safety monitoring system for construction sites.

• Built real-time PPE detection pipeline using YOLOv11 with 92% mAP accuracy
• Implemented PPO-based reinforcement learning agent for dynamic safety risk scoring
• Designed modular backend architecture for scalable video stream processing

Patent-pending intelligent healthcare monitoring system.

• Developed ML-based health monitoring system using KNN and SVM with 94% prediction accuracy
• Integrated IoT sensors for real-time patient data collection and cloud-based processing
• Filed patent for novel ambient intelligence framework for healthcare applications

Scalable conversational AI system with advanced language understanding.

• Engineered end-to-end voice assistant integrating STT, LLM, and TTS pipelines
• Optimized response latency to under 2 seconds for real-time conversational flow
• Implemented context-aware dialogue management for multi-turn conversations

Distributed system for secure legal document management.

• Architected blockchain-based legal records system using Hyperledger Fabric
• Implemented smart contracts for tamper-proof document verification and audit trails
• Built RESTful API layer for seamless integration with existing legal workflows

Real-time cross-org knowledge marketplace using Neo4j skills graph and event-driven Go backend.

• Engineered a real-time cross-org knowledge marketplace using a Neo4j skills graph (10K+ engineers, 50K+ skill edges), implementing centrality-based expert ranking to reduce stale backlog by 35%.
• Built an event-driven Go backend with Redis pub/sub enabling sub-200ms bounty routing and atomic payout guarantees via idempotent transitions.

Real-time speech analytics pipeline for Tamil dialects with adaptive AI learning.

• Built a real-time speech analytics pipeline for Tamil dialects using Whisper and achieved 85% transcription accuracy with <2s inference latency.
• Architected a scalable WebSocket-based backend with adaptive AI-generated exercises, improving efficiency by 40% via async optimization.

Scalable time-series log analysis system processing 1M+ API records to detect latency regressions.

• Built a scalable time-series log analysis system processing 1M+ API records, detecting post-deployment latency regressions using statistical anomaly detection.
• Improved detection accuracy by 30% and backend query performance by 35% via indexed schema and async processing.

High-performance client projects, custom LLM solutions, and digital automation systems.

• Designed and engineered bespoke AI integration solutions and high-performance web applications for global clients.
• Developed custom LLM integration pipelines, search indexing systems, and responsive frontends resulting in 50% improvement in user engagement.
• Managed end-to-end development lifecycle, client communication, and deployment to AWS and Vercel cloud infrastructures.

App developed by me to master Computer Science fundamentals powered by AI tutor recommendations.

• Developed a mobile application to help users master Computer Science fundamentals through AI-powered interactive learning.
• Integrated adaptive learning modules that dynamically tailor explanations and practice problems to user proficiency.
• Built optimized mobile layouts ensuring smooth performance, offline support, and native Android compatibility.