Credits
This library is assembled, not invented in a vacuum. The tracks below belong to the people who wrote them. We are grateful they put the work in public.
TensorTonic
Problems, ML math, LLM internals, research-paper labs, and the from-scratch method note come from TensorTonic, founded by Pratham Grover. Official exercises, starter code, and Plus stay on their site. Subscribe there if you want the paid lane.
Tensor-to-Tenant
The 108-week spine, gates, operating model, Forge, on-ramp plan, and learner cookiecutter are Sethu Iyer’s Tensor-to-Tenant. Official browser: sethuiyer.github.io/tensor-to-tenant.
Leetcode Darbar and YACR
The 548-slot catalog and the programming-paradigms note are aninokuma’s:
Problem statements stay on LeetCode. We only keep catalog metadata and links.
CS/AI 684
The 14-week long-context agent syllabus on this mirror is organized around public systems work, especially:
- DeepSeek-AI — V4.1-Flash report and V3 / MLA papers
- Kwon et al. — vLLM / PagedAttention
- Dao et al. — FlashAttention-2
- Vaswani et al. — Attention Is All You Need
- Williams, Waterman, Patterson — the roofline model
Every week lists the rest of its public readings. We did not reconstruct TensorTonic Plus labs for this course.
Landmark papers and official code
Gated study notes on this site are written from the public paper and the official repository. Each of those pages names the source. The labs we do name in one place:
| Work | Authors / org | Paper |
|---|---|---|
| DeepSeek-V3 | DeepSeek-AI | 2412.19437 |
| Llama 3 | Meta | 2407.21783 |
| Gemma 3 | Google DeepMind | 2503.19786 |
| GLM-4.5 | Zhipu | 2508.06471 |
| gpt-oss | OpenAI | 2508.10925 |
| GPT-2 | OpenAI | model card |
| Arcee Trinity | Arcee | 2602.17004 |
| Bennett on hypothesis choice | Michael Timothy Bennett | 2301.12987 |
Open research tracks (AlexNet, ResNet, Transformer, and the rest) credit the original papers on each page.
36-week on-ramp
These are the nine resources the on-ramp points at, in order:
| Resource | Who to thank |
|---|---|
| Search / docs / git primers | The instructors of those public videos |
| CS50x 2024 | David J. Malan and the CS50 staff at Harvard |
| Nand to Tetris | Noam Nisan and Shimon Schocken |
| MIT 6.042J | Tom Leighton, Marten van Dijk, and MIT OCW |
| YACR | aninokuma |
| Statistical Learning with Python | Gareth James, Daniela Witten, Trevor Hastie, Robert Tibshirani |
| Software Engineering at Google | Titus Winters, Tom Manshreck, Hyrum Wright |
Tools on this site
Math rendering is KaTeX, created at Khan Academy and maintained by the KaTeX authors. Progress, the desk, and evidence notes stay in your browser.
This mirror
AI Engineer Exhaustive is a study library. It does not replace TensorTonic Plus, the official T2T site, or any paper’s authors. If we missed a name you can document, say so and we will add it.