FineEnvs — RL Environments 101: building and scaling RL environments in the age of LLMs

A cross-platform GUI automation Python module for human beings. Used to programmatically control the mouse & keyboard.

Contrastive Language Models (CLMs) are a new class of System One model trained with a contrastive learning objective that connects states and actions.

Convert GitHub PRs into Harbor tasks

Bird frame for Raspberry Pi or your homelab - real-time bird detection by audio, fully local AI, rendered as real, hand-cut 1800s bird illustrations. On an e-ink panel, a TV, or any screen.

Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

RL-training knowledge-work agents on APEX-Agents data with SkyRL + Harbor

Archived. Current work: Charming (https://usecharming.com).

AgentENV (AENV) is a distributed platform for running agent environments at scale.

Synthetic pretraining data by rephrasing the web

Framework for evaluating and improving agents

Repository for results and data (coming soon!) for ClawsBench

SWE-agent takes a GitHub issue and tries to automatically fix it, using your LM of choice. It can also be employed for offensive cybersecurity or competitive coding challenges. [NeurIPS 2024]

Code and documentation to train Stanford's Alpaca models, and generate the data.

Synthetic data curation for post-training and structured data extraction

Democratizing Deep-Learning for Drug Discovery, Quantum Chemistry, Materials Science and Biology

A construction kit for reinforcement learning environment management.

DeepSeek Harness: Everything is a Plugin.

[NeurIPS 2023, Spotlight] Rank-N-Contrast: Learning Continuous Representations for Regression

Hydra is a framework for elegantly configuring complex applications

A self-improving RLM agent for coding workflows and long-running autonomous tasks.

[NeurIPS 2025 D&B Spotlight] Scaling Data for SWE-agents

MoonEP: A Perfectly Balanced Expert Parallelism Library via Dynamic Redundant Experts

Inspect: A framework for large language model evaluations

ToolOrchestra is an end-to-end RL training framework for orchestrating tools and agentic workflows.

Developer Asset Hub for NVIDIA Nemotron — A one-stop resource for training recipes, usage cookbooks, datasets, and full end-to-end reference examples to build with Nemotron models

Data recipes and robust infrastructure for training AI agents