Great code
doesn’t just execute
commands.
It learns,
adapts, and
creates solutions.
"Work that
taught me most."
Designed and deployed a production-ready voice assistant serving 30+ Indian languages with real-time multimodal processing.
Provides accessible voice technology to underserved linguistic communities with intelligent routing, multimodal processing, and culturally aware responses.
Published research in IEEE OTCON 2025, demonstrating scalable architecture for culturally inclusive conversational AI.
Saahayak
Saahayak is an AI-powered teaching assistant designed to support educators managing multi-grade classrooms. Built with Google Gemini and Genkit, it generates hyper-localized content, differentiated worksheets, visual aids, and lesson plans—all from text, voice, or image inputs in 25+ Indian languages including Hindi, Marathi, and more.
Built to address the unique challenges of rural Indian schools where one teacher manages multiple grades with limited resources. This project demonstrates how practical AI can save educators hours of preparation time while maintaining the human touch that makes great teaching possible.
Custom SLM: Personalized AI Assistant
I'm training a Small Language Model on my experiences, knowledge, and problem-solving patterns—creating an AI version of how I think, code, and approach challenges.
This isn't about replacing myself; it's about capturing my expertise in a model that can assist others even when I'm not available.
This project bridges AI engineering with self-documentation, demonstrating both technical capability and philosophical understanding.
It's a living portfolio that evolves with me, learns from my work, and serves as an accessible interface to my knowledge—showing how SLMs can be personalized tools, not just generic assistants.
The world is one big data problem.
Tools and frameworks I work with daily.
.svg)
![](_astro/svgs/Vercel.svg")
Credentials & Recognition
I build
solutions.
Featured Projects
This project is a sophisticated RAG-based system designed to answer queries from PDF documents by leveraging both text and images. It uses LangChain and Google Gemini Pro to provide accurate, context-aware responses.
View Project on GitHubThis project demonstrates how to run and manage models locally using Ollama by creating an interactive UI with Streamlit.
View Project on GitHubEver found yourself staring at a blank screen, desperate for a data project idea that actually shows off your skills? It’s a tool built for data folks like us—analysts, scientists, engineers—who want project ideas that feel fresh, relevant, and fun. Using a bit of AI magic, it takes your job title, favorite tools, and industry, then whips up personalized project suggestions complete with details, timelines, and skills graphs to help you bring them to life.
View Project on GitHubproject leverages the power of the Gemini Vision Model to create an advanced system for generating invoices in multiple languages. Whether you're managing international transactions or need versatile language support, this application provides a solution with state-of-the-art computer vision capabilities.
View Project on GitHubA comprehensive AI-powered desktop assistant that combines multiple AI models to create an intelligent programming companion. This tool integrates visual processing, text analysis, and voice interaction to enhance programmer productivity through an intuitive interface.
View Project on GitHubThe purpose of this repository is thus to provide a comprehensive and dynamic hub to showcase the latest retrieval augmented generation (RAG) systems, and to enhance their error reduction as well as processing speed and contextual richness. It’s the go-to resource for those looking to become an RAG master — an assembly of curated Jupyter notebooks for learning and teaching RAG concepts. If you’re trying your hand at retrieval-based approaches to improve AI models or are just a fungi out for some advanced RAG methods, this repository is your thing.
View Project on GitHubThis project is a comprehensive exploration of advanced techniques in building chatbots using LangChain, a powerful framework for developing applications with large language models (LLMs).
View Project on GitHubA Python-based multi-agent AI system built with the CrewAI framework for orchestrating collaborative autonomous agents that work together to solve complex tasks.
View Project on GitHubDeploy NVIDIA'S GPU Accelerated AI models as API using Langserve
View Project on GitHubThis project contains a Python script for real-time object tracking using OpenCV.
View Project on GitHubA deep learning project that uses the VGG-16 convolutional neural network for automated food image classification. The model leverages transfer learning with pre-trained ImageNet weights to accurately classify different types of food from images.
View Project on GitHubThis project enables users to generate images from voice input by combining speech recognition with high-speed text-to-image generation. The system leverages NVIDIA TensorRT optimization for real-time image synthesis, capable of generating images in under a second.
View Project on GitHubA project aims to provide detailed notes based on YouTube video transcripts across various subjects including Open-CV, Machine Learning, Large language Models, and Data Science & Statistics, Generative -AI. With the power of AI, you can now convert video transcripts into comprehensive study materials.
View Project on GitHub