AI & Cloud Engineer focused on building reliable, real-world AI systems. I develop local RAG assistants, machine-learning pipelines, and cloud-native microservices using AWS and Kubernetes, grounded in academic experience from IIT Kharagpur and Saras AI Institute.
Developed a local RAG-powered YouTube AI assistant using Llama 3.2, Ollama, and LangChain to converse with any video. Engineered a yt-dlp, nomic-embed-text, and ChromaDB pipeline for semantic search over YouTube transcripts with FastAPI REST endpoints.
Built an end-to-end churn prediction system for a 7,000+ customer telecom dataset with domain-specific feature engineering. Benchmarked Logistic Regression, KNN, and Random Forest (80.4% test accuracy) and deployed as a serialized Flask REST API on AWS EC2.
Architected and deployed a multi-tier microservices application (Flask API, Web UI, background Worker, and Redis broker) using Docker containers, Kubernetes manifests, and Helm charts on AWS EKS with AWS Elastic Load Balancing (ELB).
Performed review sentiment analysis with VADER and LDA topic modeling to isolate top passenger pain points. Built a TF-IDF + Random Forest model (RMSE 0.406) and a serverless recommendation engine on AWS Lambda and API Gateway.
Built an incremental cryptocurrency data pipeline using CoinGecko API with retry handling and JSON ingestion. Architected a Snowflake warehouse using 3NF and Star Schema modeling with automated ETL via Snowflake Streams and Tasks.
I am an AI and Cloud Engineer passionate about building pragmatic systems that solve real-world problems. My core focus spans Generative AI, Retrieval-Augmented Generation (RAG), statistical Machine Learning, and cloud-native container orchestration.
Whether benchmarking Vision-Language Models on GPUs or orchestrating distributed microservices across Kubernetes clusters, I thrive on turning complex engineering ideas into rock-solid, production-ready systems.
Enhanced an adaptive PDF processing pipeline using LiteParse and OCR/VLM benchmarking to optimize document intelligence and extraction workflows.
Solved over 300 algorithm and data structures challenges with optimal space and time complexity.
Won Silver Medal in the 2025 General Championship Nukkad (street play) representing Patel Hall dramatics.
Demonstrated daily discipline and consistency with a 600+ day Duolingo streak in Japanese (Score: 16).
Interested in AI engineering opportunities, projects, or collaboration? I'd be glad to connect and talk through technical ideas.