Education
Hindustan Institute of Technology and Science, Chennai, IndiaSep 2022 – May 2026
B.Tech. in Computer Science (Specialization: AI & Data Science)CGPA: 8.26/10
Skills
Languages & Frameworks: Python, R, JavaScript, Node.js, TensorFlow, PyTorch, Scikit-learn, NumPy, Pandas
AI/ML: Hugging Face, NLP, Federated Learning, Regression, Classification, Clustering
Cloud & DevOps: Google Cloud Platform (GCP), Vercel, Docker, Git, CI/CD
Databases: Supabase (PostgreSQL, Auth, Storage, Realtime APIs)
Academic Projects
Intelligent Query Routing Framework for LLM OrchestrationSep 2025 – Jan 2026
IEEE Conference Presentation, March 2026 | GitHub
- Designed and implemented a web-based intelligent routing system that dynamically classifies user queries and routes them to optimal language models (GPT-4, Claude 3.5, Llama 3) based on task complexity, achieving 93.7% routing accuracy.
- Developed a lightweight intent classification algorithm using heuristic-based keyword matching and complexity scoring, completing query categorization in under 20ms with minimal computational overhead.
- Built a production-ready serverless architecture using Next.js 14 and Vercel Edge Functions, demonstrating 42% cost reduction and 39% latency improvement over monolithic single-model approaches across 100 diverse test queries.
- Implemented real-time streaming responses via Server-Sent Events, security measures including PII redaction and prompt injection mitigation, achieving scalability up to 10,000 concurrent requests.
REX Healthify — AI Health AssistantOct 2025 – Nov 2025
4th Place, OpenAI × NxtWave Buildathon | GitHub
- Developed an AI-driven health assistant using Gemini AI for personalized medical guidance with features for medical record management and emergency doctor discovery.
- Built medication reminders, voice interaction capabilities, and NFC-based emergency access to critical health data for rapid response scenarios.
- Competed against 1,000+ teams and secured 4th place for building a high-impact technical solution.
Privacy-Preserving Federated Learning for Drone SwarmSep 2025 – Nov 2025
IEEE International Conference, Singapore (2025) — Best Paper Award | GitHub
- Proposed a privacy-preserving federated learning framework enabling decentralized exploration in autonomous drone swarms without centralized control.
- Designed a 3D voxel-based environment with room-like obstacles and implemented LiDAR sensor simulation using 3D spherical ray marching with 8m sensing range.
- Developed frontier-based exploration strategies and applied Federated Averaging (FedAvg) to aggregate drone models while preserving local data privacy.
- Built real-time 3D visualization tools to monitor drone trajectories, environment coverage, and swarm coordination.