A new learning platform teaches systems engineering by having developers rebuild foundational infrastructure from scratch. The curriculum spans 80+ courses across 9 languages where every lesson follows a choose-write-run loop — students pick a component, implement it, execute tests, and iterate until the behavior clicks. This mirrors how senior engineers actually learn: by confronting the same constraints that shaped the original designs.

The platform organizes content into 14 career paths with sequenced hour estimates: Backend Engineer (~80h) covers Go, SQL, caches, queues, and wire protocols; Frontend Engineer (~65h) progresses through JavaScript, TypeScript, and framework internals; Full-Stack (~90h) combines both; DevOps/SRE (~70h) tackles Bash, Go tooling, Docker, CI, and observability; Data Scientist (~60h) focuses on Python, R, SQL, and query data structures. Each language offers a 37-language Fundamentals → Intermediate → Advanced track.

The technical scope is deliberate: rebuilding Redis teaches event loops and protocol parsing; rebuilding Git teaches content-addressable storage and Merkle trees; rebuilding a database teaches B-trees, WAL, and concurrency control. These aren't toy exercises — they're the actual primitives that production systems compose.

The pedagogical bet is that constraint-driven implementation beats passive consumption. By removing copy-paste escape hatches, the platform forces the mental model formation that distinguishes engineers who debug distributed systems from those who only configure them. Free tier with no credit card requirement lowers the activation energy for teams evaluating internal upskilling.