I've been building software companies since 2008. Three startups, one exit, and I'm still working on my current startup since 2018. Countless technical decisions that seemed smart at the time but aged poorly. You know how it goes—you pick a technology stack, build everything on it, and then watch the industry pivot in a completely different direction six months later.

Machine learning hit my radar seriously around 2015, but I'll be honest—I initially dismissed it as academic research that wouldn't affect "real" software development for years. That was... not my best prediction. By 2018, I was scrambling to understand why my traditional search and recommendation systems suddenly felt archaic compared to what a few ML engineers could build in a weekend.

The 2022-2023 AI explosion caught everyone off guard, including people who should have seen it coming. Suddenly, every business conversation included the phrase "AI strategy," and I realized I had a choice: become one of those executives who makes AI decisions based on marketing materials, or actually understand what these systems do.

So I spent the better part of a year reading papers, implementing toy versions of transformers, and trying to bridge the gap between the math and the practical implications. This site is essentially my notes from that journey, organized in a way that might be useful to others facing the same challenge.

There are already excellent resources for learning AI, each serving different needs. Academic courses like Stanford's CS224N provide rigorous theoretical foundations. Coursera and edX offer structured paths with university credentials. YouTube has countless practical tutorials. The research paper ecosystem is incredibly rich if you know where to look.

Each of these has its place. I've learned from all of them. But I kept finding gaps—theory that didn't connect to implementation, tutorials that worked in isolation but fell apart in real systems, academic content that assumed mathematical backgrounds I didn't have.

This site tries to fill some of those gaps. It's not comprehensive, it's not the only way to learn these topics, and it's definitely not the authoritative source. It's one person's attempt to organize knowledge in a way that bridges theory and practice, aimed at people who need to make actual decisions about these technologies.

The content here is aimed at working professionals who need to understand AI systems well enough to make informed decisions. That includes software engineers evaluating AI tools, product managers designing AI features, technical leaders planning AI strategies, and consultants implementing AI solutions for clients.

You don't need a machine learning background, but you should be comfortable with technical concepts. I assume you've written code, understand basic data structures, and have experience with software systems. I try to explain the math when it's essential, but I don't assume you love linear algebra.

If you're a complete beginner to programming, this probably isn't the right starting point. If you're doing machine learning research, you likely need more mathematical rigor than I provide. But if you're somewhere in between—trying to bridge the gap between business requirements and technical implementation—this might be useful.

This isn't a comprehensive machine learning course. It's not a certification program. It won't make you an AI researcher. It's focused specifically on language models and related technologies, with an emphasis on understanding how they work well enough to use them effectively.

The content reflects my own learning journey, which means it has gaps. I skip mathematical proofs that I don't think are essential for practical understanding. I focus on concepts that have proven useful in real implementations. I update things when I learn better explanations or when the technology changes significantly.

I've made mistakes before, and I'll make them again. When readers point out errors or suggest improvements, I try to incorporate that feedback. The goal is practical understanding, not academic perfection.