Thank you for contributing to Awesome Agentic Engineering Resources! π
This list is a resources list, not a tools list. Open-source tools for building agentic systems belong in the sibling list awesome-production-agentic-systems; production ML tools belong in awesome-production-machine-learning. Please submit pure-tool entries to the appropriate sibling list.
Every entry should be:
- Substantive β a full article, essay, book, course, paper, cookbook, playbook, talk, podcast, benchmark, or reference implementation. Tweet threads, marketing pages, and short announcement posts do not qualify.
- Canonical β repeatedly cited by multiple independent authorities (canonical authors, course syllabi, survey papers, widely-read engineering blogs), or clearly destined to become so.
- Primary source β link to the original publication, not an aggregator or reposted copy.
- Durable β hosted on a stable domain. Avoid paywalled Medium posts, personal domains that frequently move, and link shorteners.
- Not already listed β search the README before opening a PR.
- Not a pure tool β if the resource is primarily an open-source library, submit it to the sibling tool lists instead.
This README is structured as a matrix:
- Row = the type of resource (Core & Foundations, Courses, Books, Articles, Tutorials & Cookbooks, Playbooks, Papers, Benchmarks, Reference Implementations, Talks, Podcasts, Newsletters, etc.).
- Topic (T1βT21) = the subject area (Coding Agents, RAG, Memory, Evaluation, Observability, Guardrails, and so on). The topic list is in the Topic Coverage Matrix.
Pick the single strongest-fit row and topic for your resource. Each entry belongs to one cell; cross-references are reserved for the Trending, Milestones, and Core & Foundations sections.
Entries are one-line bullet points, sorted alphabetically within their cell:
* [Resource title](https://example.com/resource) π οΈ π β One short, informative sentence that explains what the reader will learn or why it matters (author / organisation, year where relevant).Guidelines:
- Icon tags β use the icons from the Legend. Combine tags where useful (e.g.,
π§βπ π°for a paid course,π π§ͺfor a paper accompanying a benchmark). - Description β one sentence, informative not promotional. Explain why this resource matters.
- No trailing punctuation clutter β end the sentence with a single period.
- Author or org in parentheses at the end is encouraged for articles, papers, talks, and books.
- Do not include tracking parameters or affiliate links in URLs.
- GitHub repositories that double as a resource (e.g., a cookbook repo) may include a stars shield:
.
* [Building LLMs for Production](https://www.towardsai.net/book) π π° β End-to-end guide to designing, fine-tuning, and deploying LLM applications, covering RAG, evaluation, and observability (Louis-FranΓ§ois Bouchard & Louie Peters, 2024).* Check out this amazing new tool I built! β marketing, no link
* [My Blog](https://mysite.com) β Great content β not substantive, vague
* [LangChain](https://github.com/langchain-ai/langchain) β a tool, belongs in sibling listBefore opening a PR:
- The resource is substantive, canonical, primary-source, and durable.
- I searched the README and confirmed this entry is not already listed.
- I chose the strongest-fit row Γ topic cell.
- My entry follows the one-line format with appropriate icon tags.
- The link resolves (no 404, no redirect chain).
- My entry is alphabetically sorted within its cell.
- This is not a pure tool (those belong in the sibling lists).
Prefer an issue over a PR when:
- You want to propose a new topic (a new
Tcell in the matrix). - You want to nominate a resource for the Trending section.
- You found a broken link but are not submitting a replacement.
Issue templates are available under New Issue for each of these.
Please read our Code of Conduct. We expect all contributors to uphold it.
By contributing, you agree that your contributions will be released under the same CC0-1.0 terms as this repository.