Autonomous AI-powered academic research system.

Autonomous deep-research platform powered by Hermes Agent. Multi-phase pipeline with parallel research, self-critique, iterative refinement, and real-time streaming.

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

AI-Powered Learning System — Claude Code + NotebookLM + Obsidian. Role-based curriculum builder with YouTube ASR, academic papers, interactive quiz/flashcards, and Socratic study companion.

Multi-agent research pipeline for Claude Code — web crawl, knowledge graph, synthesis, and citation-rich reports via slash commands.

Beautiful modular D3QN research pipeline with training, ablations, plots, report, and packaging

Workflow manager module for the Storm platform.

A Python CLI tool that runs a multi-agent research pipeline — search, scrape, summarize, critique, and synthesize into a cited report. Built with Groq, Gemini, DuckDuckGo, and pure manual orchestration. No LangChain.

End-to-end LLM-augmented Handwritten Text Recognition (HTR) pipeline for early modern Spanish manuscripts, combining TrOCR decoding with linguistic beam reranking and orthography-preserving contextual correction.

Local-first autonomous growth engine for Parad0x-Labs: research, strategy, creative generation, approval-gated publishing, metrics, and memory in one portable, testable platform.

Sample-backed text analysis pipeline for dark-web listings and forum posts using classification and topic summaries.

Deep Research Operating System — multi-agent, multi-model research pipeline with a Padel AI Coach case study. Apache 2.0.

Production-grade PyTorch framework for computational pathology research. Features attention-based MIL models, foundation model integration (Phikon/UNI/CONCH), clinical PACS integration, and comprehensive testing (1,448 tests). Validated on PCam (85.25% accuracy, 93.94% AUC). Built for research and clinical deployment.