Privacy-First AI Agent Platform
checkingArchitecting a multi-agent AI platform for regulated verticals with ephemeral-session architecture, local LLM inference, and zero data persistence.
Data Science Student, IIT Madras
I'm a Data Science & Applications student at IIT Madras, focused on building privacy-first AI agents, scalable machine learning pipelines, and full-stack web applications.
Every project runs live on its own subdomain — click through and try it.
Architecting a multi-agent AI platform for regulated verticals with ephemeral-session architecture, local LLM inference, and zero data persistence.
Dual-FAISS RAG pipeline retrieving into a quantized Gemma-4-12B. Achieved top score (0.75893 MAP@3) on Kaggle. Tracked 47 experiments on WandB.
Engineered an ML pipeline with TF-IDF and metadata preprocessing to predict comment categories. Evaluated multiple classifiers achieving macro F1 0.8175.
Role-based recruitment application spanning 3 dashboards. Automated recurring reminder and reporting jobs with Redis and Celery.
Multi-user quiz platform featuring secure authentication, dynamic quiz management, and real-time Matplotlib-based performance tracking.
The stack I reach for, grouped by what I use it for.
Training programs and official credentials.
Feel free to look around, break things (it's okay, I'll fix it later), and if any doubt or query strikes, get in touch.