Explore the directory

A Django-based platform that lets life science researchers at a Swedish research institute serve machine learning models and host data science apps such as Shiny, Gradio, Streamlit, and Dash on a shared Kubernetes cluster.

• Inference
• Hosting

A companion robot project for the DeepLearning.AI short course on building AI assistants with on-device memory, using a camera and microphone to recognize objects, learn their names by voice, and recall when and where it saw them, entirely on the device.

• Agent memory
• Tutorials

An MCP server and CLI that hands off a coding-agent session to the next one by building a brief from git facts and test results rather than trusting the previous agent's own summary.

• Coding agents
• Agent memory

A CLI and GitHub Action that regression-tests RAG pipeline changes against a baseline using paired bootstrap significance tests, blocking a pull request only when a retrieval metric change is statistically real rather than noise.

• RAG
• Testing

A Python packaging CLI (pypm) bundled with a companion C memory library, PMLL, and an MCP memory server, providing hermetic package bundles, plugins, and signing alongside durable structured memory for agent workflows.

• MCP servers
• CLI tools

A protocol and CLI that moves a live conversation from one AI coding agent to another -- Claude Code, Codex, Cursor, and others -- without copying the transcript or re-writing the brief for the new tool.

• Coding agents

A Kubernetes-native workflow engine that executes AI-assisted operations as a directed acyclic graph of deterministic steps, calling an LLM only for the analysis step rather than giving an agent open-ended cluster access.

• Containers
• Workflow automation

A workflow orchestration platform where AI coding agents such as Claude Code or Cursor build and deploy Python workflows for you, with LangGraph powering the underlying execution rather than a proprietary DSL.

• Agent frameworks
• Workflow automation

A background orchestration service that routes GitHub issues to AI coding agents such as Claude, Codex, OpenCode, Kimi, or MiniMax, running each in an isolated git worktree through to an automatically opened pull request.

• Coding agents
• CI/CD

A visual, node-based editor for composing PyTorch neural network architectures as typed, executable cards, with two-way sync to generated code, step-by-step tensor execution, and a constrained AI planner that proposes only valid graph changes.

• Models
• Design tools

A visualization tool that inspects the internals of open-source LLMs the way an MRI images a brain, mapping activations, weight distributions, information flow, and anomalies for a given model and prompt across five scan modes.

• AI evals
• Debugging

A hands-on workshop that builds .NET multi-agent applications with the Microsoft Agent Framework, Microsoft Foundry, and Aspire, covering sequential, concurrent, handoff, and group-chat orchestration patterns.

• Agent frameworks
• Tutorials

A 34-chapter book for Java developers, built as a Jupyter Book static site, teaching machine learning through practicing local inference and fine-tuning of open-source GGUF models.

• Tutorials
• Books

A self-hosted MCP server that exposes the full Mistral AI API -- chat, OCR, Voxtral audio transcription, Codestral code completion, vision, batch, and durable workflows -- to Claude Code, Cursor, Zed, Windsurf, and other MCP clients.

• MCP servers

An educational, from-scratch PyTorch implementation of Mamba, the selective state space model architecture, covering the selective scan mechanism, a Mamba block, and a complete language model built on it.

• Models
• Tutorials

A restrained Python framework for building AI agents, covering tools, sessions, events, context compaction, and serving with a small set of composable primitives rather than a large platform.

• Agent frameworks

A research playground that guides an Isabelle/HOL theorem prover with large language models, combining LLM-proposed tactics with Isabelle's own nitpick, quickcheck, and sledgehammer, and a beam search over proof outlines.

• Papers

An open-source framework that benchmarks GPT-4, Claude, Gemini, Mistral, and Llama side by side on the same prompts, measuring accuracy, latency, per-token cost, a hallucination signal, and reasoning quality.

• AI evals

An open-source cognitive architecture, written mostly in Python with an optional Rust core, that models perception, memory, and causal reasoning without routing every step through a large language model, meant to run alongside an LLM agent such as Hermes.

• Agent frameworks
• Agent memory

Runnable end-to-end examples for the Kakunin AI agent identity and compliance API, showing X.509 certificate issuance and scope-enforced tool calls integrated with LangChain, CrewAI, AutoGen, LlamaIndex, Next.js, and the Vercel AI SDK.

• Auth
• Tutorials

A cross-platform Flutter chat app that connects to hundreds of AI models through OpenRouter and to remote MCP servers over Streamable HTTP, running an agentic tool-use loop against them.

• MCP servers
• Inference

A set of small, reviewable control loops for Hermes Agent installations: one that proposes and tests repository changes in a propose-test-keep-or-revert cycle, and read-only loops that audit a Hermes install's health and suggest improvements for human review.

• Agent frameworks
• Agent skills

A self-hosted AI development workspace where a user describes an application in chat, reviews the generated specification, and builds it in an integrated editor with a live preview, producing a React frontend with an optional Go backend.

• Coding agents
• Web frameworks

A retrieval-augmented generation agent over the complete 3GPP Release-18 and Release-19 5G specification set, answering natural-language questions with paragraph-level citations back to the original text and strict grounding.

• RAG