Featured Projects
A showcase of my engineering projects spanning AI/ML, data engineering, and full-stack systems. Not all repositories are public — private ones are marked accordingly.
JARVIS-style personal assistant with text and voice interaction and system-level control, built on the Claude Agent SDK. Runs fully local speech-to-text and text-to-speech, wake-word activation, persistent conversation memory across sessions, and a FastAPI backend for a future mobile client — all gated behind strict permission confirmation for anything that touches the system.
Fitness platform that logs macros from a food photo or barcode scan and generates personalized workout programs through an LLM coach grounded in a research corpus of exercise-science literature (RAG over a curated set of studies and workout data), rather than generic templates.
AI-native sensitive data governance system that ingests, cleans, and standardizes enterprise datasets, automatically detecting PII with spaCy/Presidio NLP and applying configurable tokenization and cryptographic masking rules.
Autonomous database migration engine providing zero-downtime replication to Google BigQuery via Debezium CDC and Kafka, translating legacy PL/SQL and T-SQL logic into standard SQL using AST parsing and LLM neural fallbacks.
AI-powered tabular analytics platform utilizing a Text-to-Code paradigm (Llama-3.3-70b/Groq) to compile and execute Python code in a secure sandboxed environment with a self-healing retry loop.
AI-powered semantic validation pipeline to detect documentation drift and ensure alignment between software documentation and source code within CI/CD workflows.
Cloud-native urban logistics analytics pipeline to optimize food delivery operations, identify high-demand zones, and improve metropolitan delivery intelligence.
Hybrid Retrieval-Augmented Generation platform integrating graph databases, semantic search, and LLMs for enterprise knowledge management.
Voice-controlled web platform leveraging LLMs and real-time stock data to democratize financial insights for Indian retail investors.
Natural language interface using FLAN-T5 that generates production-ready FastAPI backend code from plain English prompts.
Supply chain for safe distribution of medicines using Blockchain and AI technologies.
Smart backpack integrating RFID technology with Flutter mobile application for automated inventory tracking and real-time notifications.
Intrusion detection system built using hybrid model approach, combining supervised and unsupervised techniques for enhanced accuracy.
AI-driven system to enhance threat detection and response through advanced analytics and real-time automated interventions.
Full-stack AI application that dynamically generates FastAPI endpoints from natural language prompts with interactive dashboards.
Applied supervised and unsupervised ML techniques to classify Spotify songs by genre using audio feature data.