Matt Dancho Lists 10 GitHub Repos to Become an AI Engineer in 90 Days
On September 19, Matt Dancho, founder of Business Science and a data scientist, shared a "become an AI engineer in 90 days" learning path in a tweet thread. His advice: skip the paid courses and instead build projects directly with 10 GitHub repositories. His core logic is that the skills these repos cover already appear heavily in job descriptions, and hands-on projects beat grinding through courses.
Confirmed
- #1 on the list is LangChain, which the author says appears in nearly every AI engineer role.
- #2 is LangGraph: a lower-level orchestration framework for building long-running stateful agents as graphs; the author says the "agentic workflows" in JDs refer to repos like this.
- #3 is LlamaIndex: the de facto standard framework for RAG and document agents, with about 52.2k GitHub stars, named in almost every retrieval-pipeline JD.
- #4 is CrewAI: an open-source Python multi-agent orchestration framework that organizes agent teams by roles and tasks (34.7k stars); the author says it's already used in production by Fortune 500 companies.
- #5 and #6 are Qdrant, a production-grade vector database written in Rust, and Ragas, a RAG evaluation framework: the latter quantifies hallucination, faithfulness, relevance and more, and can even generate production-aligned test data when no test set exists.
- #7 is Ollama: run open-source LLMs locally with a single command, with 181k stars; its README shows support for models like Kimi and GLM. The author says JDs ask for local inference experience due to cost and privacy concerns.
- #8 is punkpeye's awesome-mcp-servers (95.3k stars), an index of production-ready MCP servers; the author calls MCP the newest skill appearing in current job descriptions.
- #9 is Shubhamsaboo's awesome-llm-apps: 139k+ stars and 20.4k forks, collecting 100+ open-source AI agents, Agent Skills, and RAG apps spanning single agents, multi-agent teams, voice agents, and MCP apps, all Apache-2.0 licensed and runnable real code.
Why it matters
- The list turns the "projects over courses" self-learning philosophy into concrete, actionable steps, offering direct reference value for learners pivoting into AI engineering.
- The list repeatedly uses "appears in JDs," "de facto standard," and "used in production by Fortune 500" as filtering criteria, reflecting real skill demand in today's AI engineering job market: orchestration frameworks, RAG, evaluation, local inference, and the MCP ecosystem are the key directions.
- The repos involved generally have very high star counts (Ollama at 181k, awesome-llm-apps at 139k, awesome-mcp-servers at 95.3k), indirectly confirming strong community momentum behind these directions.
2026-09-18 ~ 2026-09-19 · 11 related posts
Primary sources
- [source] awesome-llm-apps: 100+ open-source AI agents and skills with 139k GitHub stars — Saboo_Shubham_ · 2026-09-18
- [source] Want to become an AI engineer in 90 days? Build from these 10 GitHub repos, not courses — mdancho84 · 2026-09-19
- LangGraph ranked #2 on 10-repo AI engineer list, powering agents at Klarna and Replit — mdancho84 · 2026-09-19
- LlamaIndex ranked #3 on 10-repo AI engineer learning list with 52k GitHub stars — mdancho84 · 2026-09-19
- CrewAI lands at No.4 on AI engineer's 10-repo learning list, with 58.8k GitHub stars — mdancho84 · 2026-09-19
- CrewAI runs multi-agent teams in Fortune 500 production; Qdrant powers vector search in Rust — mdancho84 · 2026-09-19
- Ragas is the standard framework for measurable RAG evaluation: hallucination, faithfulness, relevancy — mdancho84 · 2026-09-19
- Ollama hits 181k stars: run open-source LLMs locally in one command — mdancho84 · 2026-09-19
- Awesome MCP Servers: a 95k-star index of every production MCP server — mdancho84 · 2026-09-19
- Awesome LLM Apps ships 100+ open-source RAG and agent templates with 139k GitHub stars — mdancho84 · 2026-09-19
1 near-duplicate retellings: mdancho84