CodeCrafters makes developers better engineers by having them rebuild Redis, DNS servers, and grep from scratch.
ENTRY ANGLES
Hands-on learning platforms designed for practitioners rather than beginners · Continuous skill development for professionals facing rapid knowledge updates · Premium learning infrastructure for advanced professionals in knowledge-intensive fields
VERTICALS
CAPABILITIES
Understanding of professional workflows and advanced practitioner needs, Ability to create high-quality, rapidly-updated technical content, Premium pricing and monetization strategies for high-earning professionals
CODECRAFTERS FOUNDER
“experienced engineers don't need structured learning”
CodeCrafters offers developers a way to become dramatically better engineers – but the method is unconventional. Instead of reading tutorials or watching videos designed for beginners, it asks developers to build a fully functional clone of a well-known piece of software.
Want to understand TCP? Build your own Redis clone or HTTP server. Want to internalize DNS? Write a working DNS server from scratch. Want to really understand regular expressions? Implement your own grep. Binary trees? Build a SQLite clone.
The same framework works for learning new languages: pick a software clone as the project, build it in the language you want to learn. Redis alone is available in Rust, Go, C, Haskell, Kotlin, and PHP, among others.
Each cloning project is broken into stages with difficulty ratings, so developers can gauge the overall scope before starting. The first two steps of any project are deliberately easy – designed to get the developer into a working state quickly and create early momentum.
Starting is frictionless: a single command scaffolds a repository on the developer's machine with the opening section of commented-out starter code. Uncommenting it and hitting run produces a first working artifact. For the Redis clone, for instance, this means a server that immediately starts listening on port 6379 – and the platform's test runner sends a packet to confirm it's alive.
From there, developers write the code themselves. The platform provides structured support along the way: annotated source code from the original software with AI-generated explanations, detailed write-ups of what needs to be implemented at each stage with worked examples, and community features that let developers see how others tackled the same step in the same language or ask questions of peers working in parallel.
The AI assistant can also field questions directly, generating synthesized answers based on how previous developers approached the same problem across different implementation choices.
The platform launched as fully free, but after mentorship from Y Combinator advisors, CodeCrafters introduced paid plans starting at $30/month for individual subscribers and $39/month per user for team plans. A free tier remains, limited to the "project of the month" and recently added projects currently in beta.
CodeCrafters was founded in 2022 and went through Y Combinator that year, receiving an initial $500,000 in funding. The startup has now closed a new round of $1.8 million.
As it turns out, building software clones as a learning method isn't unusual among experienced engineers – it's just rarely formalized into a product. CodeCrafters' website features a video of Optimizely's CTO describing how he learned Rust and internalized the Git protocol by writing his own Git implementation from scratch.
The investors behind the new round underscore this point: they include an Instagram co-founder and former CTO (now Chief Product Officer at Anthropic), the Dropbox co-founder and CTO, and several other founders and technical leaders from well-known startups. These are people who learn exactly the way CodeCrafters teaches – and they were willing to put money behind it.
Look at the landscape of programming education and the problem becomes obvious: nearly everything is aimed at beginners. That's understandable – the demand for new developers is enormous, and onboarding more people into the field is a legitimate goal. But it means experienced engineers are largely underserved when they need to learn something new.
And the assumption that "experienced engineers don't need structured learning" doesn't survive scrutiny. It's likely that most developers have never finished a programming book cover-to-cover – because those books are written for people entering the field, not people already operating at a high level.
The underlying difference is how experts and novices process new information. Novices lack mental frameworks, so they experience new knowledge as unconnected facts. They need everything unpacked – the principle, worked examples, each step broken down, repetition until the framework forms.
Experts already have rich mental models. When they encounter something new, they immediately pattern-match it against what they know, slot it into the right category, and integrate it into their existing framework. What they need isn't hand-holding – they need exposure to the new elements, with targeted support only where genuine novelty appears.
A classic chess experiment illustrates this cleanly: when chess masters, intermediate players, and beginners were each shown a board position for five seconds and asked to reconstruct it, masters recalled the most pieces – but only when the position reflected a real game state. When pieces were arranged randomly, masters performed no better than beginners. Expertise is pattern recognition, and pattern recognition only activates on meaningful structure.
CodeCrafters is essentially building on this insight: give experienced developers a real project with genuine structure, provide support only at the genuinely novel steps, and let their existing expertise do the rest. It's a very different product from anything aimed at beginners – and it's the right product for a very large, underserved audience.
The same logic applies far beyond software development. Nearly every professional education market skews heavily toward entry-level learners. But every beginner who sticks with a field eventually becomes a practitioner – and once they reach that level, the available learning infrastructure largely stops serving them.
The current moment amplifies this gap. The pace of new knowledge in technology, AI, and adjacent fields has accelerated dramatically. Professionals who could once spend years working with a stable skill set now face continuous, rapid updates. The Red Queen applies: just to maintain their position, they need to learn constantly. To advance, they need to learn faster.
Platforms designed for how professionals actually learn – not how beginners learn – are a real market opportunity. CodeCrafters is one example in software engineering, but equivalent platforms could exist across medicine, law, finance, design, and any other knowledge-intensive field where the gap between beginner resources and professional development needs is large.
The economic case is also straightforward: professionals earn more than beginners. A platform solving a genuine learning problem for senior practitioners can charge accordingly – and justify it.