amitshekhariitbhu/100-days-of-ai — explained in plain English
Analysis updated 2026-05-18
Follow a structured daily study plan to learn AI and machine learning concepts from scratch.
Use the built-in revision days to review and reinforce earlier lessons before moving on.
Share daily progress publicly to stay accountable while learning AI fundamentals.
| amitshekhariitbhu/100-days-of-ai | 0xh4ku/manga-pdf-to-epub | allstarswc/allstars | |
|---|---|---|---|
| Stars | 60 | 60 | 60 |
| Language | — | Python | TypeScript |
| Setup difficulty | easy | moderate | hard |
| Complexity | 1/5 | 2/5 | 4/5 |
| Audience | general | general | general |
Figures from each repo's GitHub metadata at analysis time.
100 Days of AI is a self-guided learning plan, not a piece of software, built to teach one AI concept a day for 100 days. It was put together by Amit Shekhar, the founder of a platform called Outcome School, and each day links out to a short blog post that explains one concept. The structure follows a fixed set of rules. A person is meant to read one linked blog post a day, spend roughly 30 to 60 minutes on it, and then write a short note in their own words to help the idea stick. The plan also encourages sharing daily progress publicly on X or LinkedIn with a specific hashtag, as a way to stay accountable. If someone misses a day, the guidance is to keep going from where they left off rather than starting over. Every so often a day is set aside purely for revision, with no new material, just time to revisit and rewrite earlier notes. The 100 days are broken into phases that build on each other. The plan starts with the basics of machine learning, like supervised versus unsupervised learning, regression, and loss functions. It then moves into how neural networks actually learn, covering gradient descent, backpropagation, and the hardware involved, like GPUs and TPUs. From there it goes deep into how modern large language models work, walking through tokenization, several kinds of attention mechanisms, and the full transformer architecture piece by piece. A later phase covers other kinds of generative models beyond language models, including diffusion models, GANs, variational autoencoders, and multimodal AI. Based on the topics listed, later phases continue into training and fine-tuning language models, and the project description mentions the plan eventually reaches AI agents in production, though the full list of remaining days is not shown in what is provided here. There is no code to install or run. This project is entirely a curated set of links to external blog posts on Outcome School's website, organized into a day-by-day schedule, plus a short set of study habits meant to keep a learner consistent over the full 100 days.
A 100-day, one-concept-a-day self-study plan that walks from machine learning basics to transformers and AI agents in production.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.