Projects Recently Starred by Experts

Starred 11 hours ago by

Didier Lopes(Founder of OpenBB). Also starred by

Jiayi Pan(Author of SWE-Gym; MTS at xAI),

Pawel Garbacki(Cofounder of Fireworks AI),

Nir Gazit(Cofounder of Traceloop), and

12 more.

spinningup by openai

0.1%

12k

Educational resource for learning deep reinforcement learning

Introduces RL terminology, algorithms, and basic theory.
Offers standalone implementations of key algorithms.
Provides curated list of important papers by topic.

Created 7 years ago

Updated 1 year ago

Starred 15 hours ago by

Zack Li(Cofounder of Nexa AI). Also starred by

Alex Chen(Cofounder of Nexa AI) and

Paige Bailey(DevRel Lead at Google DeepMind).

GenieX by qualcomm

0.4%

Nexa SDK: local inference framework for GGML/ONNX models

Supports text, image, audio, and multimodal model inference.
Compatible with CPU, GPU (CUDA, Metal, ROCm, SYCL), and iOS devices.
Offers an OpenAI-compatible API server with function calling.
Includes a Streamlit UI for interactive model deployment/testing.

Created 1 year ago

Updated 8 hours ago

Starred 19 hours ago by

Wing Lian(Founder of Axolotl AI). Also starred by

Victor Taelin(Author of Bend, Kind, HVM).

llm-as-a-verifier by llm-as-a-verifier

2.7%

417

A framework for fine-grained LLM verification and trajectory reward modeling

Scales LLM evaluation with granular scoring, repeated checks, and decomposed criteria for enhanced accuracy.
Leverages Gemini's logprob extraction and a novel reward formula for trajectory assessment.
Employs round-robin tournaments to select optimal trajectories based on probabilistic rewards.
Achieves state-of-the-art results on complex agent benchmarks like Terminal-Bench and SWE-Bench.

Created 2 months ago

Updated 18 hours ago

Starred 23 hours ago by

Didier Lopes(Founder of OpenBB). Also starred by

Elvis Saravia(Founder of DAIR.AI),

Thomas Wolf(Cofounder of Hugging Face), and

Maxime Labonne(Head of Post-Training at Liquid AI).

EQ-Bench by EQ-bench

0.5%

439

Benchmark for emotional intelligence evaluation in LLMs

Assesses creative writing via LLM-as-judge using detailed criteria.
Includes Judgemark, evaluating a model's ability to judge creative writing.
Supports llama.cpp, OpenAI-compatible APIs (e.g., Ollama), and Oobabooga.
Features result uploads to Firebase and Google Sheets.

Created 2 years ago

Updated 1 year ago

Starred 1 day ago by

Shizhe Diao(Author of LMFlow; Research Scientist at NVIDIA). Also starred by

Vasek Mlejnsky(Cofounder of E2B) and

Theo Browne(Founder of Ping.gg).

Stirrup by ArtificialAnalysis

1.2%

485

Lightweight framework for building flexible AI agents

Enables agents to leverage web search, code execution, and document processing for task completion.
Supports multimodal inputs including images, video, and audio with automatic format conversion.
Offers customizable tool definition and flexible LLM provider support via OpenAI-compatible APIs and LiteLLM.
Prioritizes model-centric workflows, avoiding rigid structures that can degrade agent performance.

Created 7 months ago

Updated 14 hours ago

Starred 1 day ago by

Clement Delangue(Cofounder of Hugging Face). Also starred by

Jeff Hammerbacher(Cofounder of Cloudera),

Abubakar Abid(Cofounder of Gradio), and

Julien Chaumond(Cofounder of Hugging Face).

tau by huggingface

81.6%

819

Coding agent for terminal development and learning

Facilitates code analysis, editing, and execution using built-in tools and a TUI/CLI interface.
Integrates diverse LLM providers (OpenAI, Hugging Face, local) via a provider-neutral event stream.
Offers durable, inspectable JSONL sessions with resume and branching for persistent context.
Serves as a readable, minimalist example for understanding and building coding agent systems.

Created 3 weeks ago

Updated 8 hours ago

Starred 1 day ago by

Dror Weiss(Cofounder of Tabnine).

openalgo by marketcalls

2.5%

Algo trading platform for seamless broker integration

Connects to major trading platforms like Tradingview, MetaTrader, and Excel.
Provides real-time market data via WebSocket and a unified RESTful API.
Features an API analyzer for risk-free strategy testing and validation.
Integrates with AI assistants via Model Context Protocol for natural language trading.

Created 2 years ago

Updated 1 day ago

Starred 1 day ago by

Yiran Wu(Coauthor of AutoGen).

mellea by generative-computing

169.4%

Orchestrate complex AI workflows with generative programming

Build structured, maintainable, and efficient AI applications using opinionated prompting and sampling.
Features LLM-based verifiers and sampling for robust output validation and quality control.
Integrates with diverse inference providers and model families, supporting LoRA for efficient checks.
Enables LLM integration into legacy code and defines functions via generative slots.

Created 11 months ago

Updated 18 hours ago

Starred 1 day ago by

Abubakar Abid(Cofounder of Gradio).

neuronpedia by hijohnnylin

5.5%

985

Open-source interpretability platform for neural networks

Explore model activations, steering, and auto-interpretations with a web UI.
Supports local inference, custom dashboards, and data import via PostgreSQL.
Integrates Delphi for automated feature explanations and scoring.
Enables semantic search of explanations using OpenAI embeddings.

Created 3 years ago

Updated 11 hours ago

Starred 1 day ago by

Kevin Hou(Head of Product Engineering at Windsurf). Also starred by

Travis Fischer(Founder of Agentic).

OfficeCLI by iOfficeAI

17.6%

10k

AI-native CLI for Office document automation

Enables AI agents to read, edit, and automate Word, Excel, and PowerPoint files programmatically.
Offers structured JSON output and path-based addressing for deterministic, AI-friendly document manipulation.
Features a three-layer architecture (Read, DOM, Raw XML) and live preview for iterative AI-driven document design.
Deploys as a single, zero-dependency binary, simplifying integration into headless and CI/CD environments.

Created 3 months ago

Updated 13 hours ago

...