ContextGem: Effortless LLM extraction from documents

Karpathy’s LLM Wiki, 100% local with Ollama. Drop Markdown notes → AI extracts concepts → your Obsidian wiki auto-links and grows. Zero sharing. Your notes stay yours.

Document Summarization App using large language model (LLM) and Langchain framework. Used a pre-trained T5 model and its tokenizer from Hugging Face Transformers library. Created a summarization pipeline to generate summary using model.

A powerful CLI tool for extracting text from documents using DeepSeek OCR and generating high-quality datasets with LLM assistance.

Trade-Alert is a notification system that keeps users updated on critical news impacting their stock portfolios. It simplifies staying informed by delivering timely notifications for important articles, eliminating the need to monitor multiple platforms in today’s fast-paced market.

This Project, I worked on the development of an LLM-powered AI chatbot using Gemini 2.5 Flash, LangChain, Streamlit, and LangSmith observability. While building the system, I analyzed the LLM run logs to better understand how prompts flow through the pipeline and how responses are generated

Compose, train and test fast LLM routers

AI-powered local restaurant researcher. Finds high-quality spots using Gemini + Google Maps grounding, applies strict filters, adds value scoring, and emails clean HTML reports. Built with Streamlit.

This project implements a pipeline for training a GPT-style Language Model (LLM) from scratch using PyTorch and the Hugging Face Transformers library.

An LLM-powered system that converts natural-language requirements into validated PlantUML behavioral diagrams through a multi-stage pipeline LATO, with real-time streaming and interactive canvas rendering.

A type-safe graph execution framework built on top of OpenLit for LLM pipelines

My personal scientific research agentic codebase and documents.

CLI tool for LLM prompt pipelines. Reusable. Shareable. Scriptable.

AI-powered hiring assistant that automates end-to-end candidate evaluation using a RAG pipeline, FAISS-based semantic retrieval, and LLM-driven question generation and scoring.

This project focuses on fine-tuning Meta’s LLaMA 2 model to develop a domain-specific medical chatbot capable of understanding and responding to patient and clinician queries with high accuracy. Leveraging parameter-efficient fine-tuning techniques—LoRA and QLoRA the project ensures resource-efficient training while maintaining high performance.

An autonomous multi-agent CrewAI system that hunts for exoplanets in raw NASA telescope data using real astrophysics and rigorous signal processing.

Governed, reproducible pipeline for LLM-generated radiology board exam learning content. Companion code for Paper 1 (Lancet Digital Health) and Paper 3 (BMC Medical Education).

Sage – Prompt-Based Data Generation & Annotation Platform

Turn messy survey responses into clean research insights. Dual-model pipeline: Claude Opus 4.5 extracts themes and assigns participants, GPT-5.1 writes executive summaries. Tuned temperatures for precision where it matters.

This project aims to build an AI-powered Legal Advisor that leverages natural language processing and vector search technology to provide users with legal guidance based on authoritative legal texts.