AI-powered resume screening platform built with Spring Boot 3, MongoDB, Elasticsearch, Kafka & Python NLP (spaCy + sentence-BERT). Parses resumes, extracts skills, scores job matches, and surfaces top candidates to recruiters — all asynchronously. One docker-compose up to run.

Package vectors, chunks, configs, and retrieval memory into portable executable .rag artifacts.

Atlas-AI is a research-grade local AI system. It implements a custom inference engine using llama-cpp-python for running quantized GGUF models and sentence-transformers for local embeddings, offering full control over the RAG pipeline without external runtime dependencies.

Semantic grouping of archived emails built on top of the local email archive data from msgvault. This uses unsupervised learning for the unstructured data of Emails to showcase the emerging pattern in them,

NLP-based semantic clustering and similarity search system for national anthems using Sentence-Transformers and KMeans.

Self-evaluating RAG benchmark over SEC 10-K filings. 12 configurations evaluated with RAGAS, MLflow, and a hybrid retriever with cross-encoder reranking.