An accessible, reliable, and efficient platform for medical information and support using LLMs

The source code used for paper "Scientific Paper Retrieval with LLM-Guided Semantic-Based Ranking", published in Findings of EMNLP 2025.

Multi-user AI orchestration backend for Custom GPT and agent integrations on Azure Functions.

Structure-first pipeline that transforms GitHub repositories into question-indexed, graph-grounded semantic assets for code search, Graph-RAG, and LLM-based reasoning.

You're mid-session and your thinking shifts. Every tool saves what you said. None of them mark where it turned. Savepoint marks cognitive waypoints so you or the model can find the way back in. v3.1, open source.

Fine-grained semantic indexing of biomedical literature (a Weakly-Supervised approach)

Official repository for AIO.CORE: The autonomous protocol for AI Optimization (AIO). Eliminates Semantic Drift and Context Fragmentation in LLMs through high-density vector indexing.

Canonical protocol, deliverable schemas, and audit artifacts for RAG-ready semantic indexing (v4.23).

Dual FAISS indices: semantic (text) + structural (primitives). Segment registry, query CLI. Behavior-first retrieval combining interaction mechanics + semantic similarity.

Fine-grained semantic indexing of biomedical literature