By peter_lzh 4 Powerful Open-Source AI Projects Worth Trying
We dug up four genuinely useful open-source AI projects — spanning AI-assisted job hunting, value-investing workflows, a desktop robot cat, and a personal research dashboard. Each one is open-source and ready to fork.
01 · ai-job-search
ai-job-search is an AI job-hunting framework built on Claude Code. Its author — a geophysicist — used it to land an AI engineer offer. The core idea: you fork the repo, fill in your profile, then chain commands such as /setup, /scrape, /rank, and /apply into a full pipeline that scores role fit, and uses draft-review dual agents to customize a LaTeX résumé and cover letter and even prep you for interviews.
It ships with a built-in Danish job-portal search and can be extended with templates for your local job boards. The emphasis is on high-quality, personalized applications rather than blind mass applying.
GitHub: MadsLorentzen/ai-job-search
02 · ai-berkshire
ai-berkshire encodes the value-investing methodology of four masters — Buffett, Munger, Duan Yongping, and Li Lu — into an AI Skill compatible with both Claude Code and Codex. Rather than a vague “should I buy this?” chat, it forces a structured pass / fail / gray-zone decision and includes an adversarial “four-masters debate” mechanism.
It supports /investment-team to launch four independent agents researching the same target in parallel. All calculations use Python’s decimal.Decimal for exact precision, and key data is dual-source verified. It is a helper for building your own systematic research workflow — not a tool that makes decisions for you (investing carries risk).
GitHub: xbtlin/ai-berkshire
03 · AI_DesktopCat_Qwen3.5Omni
AI_DesktopCat (AICat) is an adorable desktop AI robot cat built on the Seeed XIAO ESP32S3 Sense. It integrates an ST7789 mini-screen for expressions, an onboard camera, a microphone, a MAX98357A speaker, servo motion, and web control into one package.
The backend runs a model such as Qwen3.5 Omni for smart conversation. The 3D-printable model, the hardware/software code, and the wiring guide are all open source, so hardware enthusiasts can replicate and mod it step by step as a moving, talking AI companion on the desk. One caveat from the author: keep the servo power supply on a common ground so you don’t fry the board.
GitHub: AI-FanGe/AI_DesktopCat_Qwen3.5Omni
04 · Vibe-Research
Vibe-Research is a personal AI investment-research dashboard covering A-shares, US stocks, and HK stocks. It does not make buy/sell decisions for you; instead it gathers market data, research reports, valuations, financials, announcements, capital flows, and news into one clean interface, and leaves an open API for you to plug in your own LLM.
Features include a daily review, a news radar (12 sectors, 108 public sources), quick stock-data views, a sector center, a watchlist, and holdings records. Data comes from public sources such as Tencent and Eastmoney, and the direction and conclusions of any AI analysis are decided entirely by the model you configure — ideal if you want a local, objective-data-plus-AI-Q&A environment.
GitHub: simonlin1212/Vibe-Research