embeding

GLiSE: Grey Literature Search Engine

Interactive Streamlit app for exploring text embeddings — upload text, generate vectors with Sentence Transformers, reduce dimensions with UMAP/t-SNE, cluster with KMeans, and visualize semantic relationships in 2D.

word embeddings. Created Date: 12 Feb 2019

app that uses Ollama embedding model and llm model to to optimize the response of the gpt via using outside sources

Agentic AI Internship Workspace

AI-powered system for detecting text conflicts and circular inconsistencies. Utilizes LLMs, NLI, BERT embeddings, and a vector database. Includes a Python backend API and UI, fully containerized with Docker.

GPT2 From Scratch using PyTorch. Inspired by Andrej Karpathy's implementation of microGPT

Production-ready AI chatbot with persistent memory (STM + LTM), built on Redis Stack for vector search (RAG), caching, and session storage. Uses Google Gemini for chat and embeddings, FastAPI backend, and a Streamlit frontend with automatic cookie-based user identity — remembers users even weeks later, no login required.

Word embedding via tensor decomposition.

Semantic document retrieval system, exposed via MCP, for AI agents like Claude

resume ranker using azure openai embedding model and using azure openai llm to analze it