Computer Science (AI & ML) undergraduate building Machine Learning and AI applications, with a focus on deep learning, Generative AI, and practical AI systems. I also contribute to open source and enjoy building projects from model development to deployment.
Languages: C++, Python, JavaScript, TypeScript, Java, C
Machine Learning: PyTorch, scikit-learn, NumPy, Pandas, Deep Learning, Transfer Learning
Generative AI: LLMs, RAG, LangChain, Vector Databases, AI Agents
Frontend: React, TypeScript, Tailwind CSS, Redux, Radix UI, Framer Motion
Backend & APIs: FastAPI, REST APIs, WebSockets
Cloud & Deployment: Google Cloud Platform, Vercel
Tools: Git, GitHub, Postman, Figma, Vite, Zod
A self-supervised and uncertainty-calibrated deep learning system for predicting the Remaining Useful Life (RUL) of aircraft engines using NASA C-MAPSS FD002.
Highlights
- Built a GRU-based RUL prediction pipeline using PyTorch
- Applied self-supervised learning with masked reconstruction and contrastive learning
- Implemented few-shot adaptation for unseen operating conditions
- Added MC Dropout-based uncertainty estimation and prediction intervals
- Evaluated model robustness under sensor noise, missing values, and sensor dropout
- Built an interactive dashboard for predictions, uncertainty, and RUL trajectories
Tech Stack
PyTorch • GRU • Self-Supervised Learning • Contrastive Learning • MC Dropout • Streamlit
Contributed responsive UI improvements, documentation, and bug fixes to the p5.js ecosystem, focusing on accessibility, responsiveness, and developer experience.
- Advanced Deep Learning
- LLM & RAG Systems
- AI Agents
- MLOps & Model Deployment
- System Design
