Systems /
Solutions.

Voice-first multilingual AI assistant helping Indian citizens access government schemes through natural conversations in Hindi & English. Built on a serverless AWS architecture with RAG-powered intent resolution across 60+ schemes.

Data analytics framework that transforms anonymized Aadhaar enrolment data into societal intelligence — detecting migration patterns, employment transitions, and demographic shifts using clustering, time-series forecasting, and a custom Life-Event Probability Index (LEPI).

Production-ready MLOps pipeline on Databricks — covering data processing, feature engineering, model training & registration, automated deployment, real-time monitoring, and CI/CD with GitHub Actions using Databricks Asset Bundles.

Full-stack predictive maintenance system with ML-driven risk scoring. Generates synthetic telemetry, predicts 30-day failure probabilities via Logistic Regression, and surfaces prioritized maintenance schedules through a React dashboard.

High-performance FastAPI backend for Sahayak AI featuring local RAG engine with ChromaDB, PostgreSQL database, voice services for STT/TTS, and Dockerized deployment with 150+ seeded government schemes.

Infrastructure sandbox for experimenting with Databricks Asset Bundles — exploring deployment pipelines, workspace configuration, and CI/CD automation for data engineering workflows.

Retrieval-Augmented Generation chatbot for HR policy queries. Uses vector embeddings to find relevant policy documents and generates contextual responses for employee questions.