Reading-first modules grounded in the Solana monorepo. Each segment is one sitting (about 10–25 minutes) unless noted. Level 1 is three reading modules only. Levels 2 and up follow the same cadence as Hyperscale: two learning modules plus one OSS contribution module (purple card) per level—hands-on work tied to real issues or mergeable slices.
Point scripts at your clone (see scripts/solana-repo.config.js, default ../other-projects/solana). After pulling upstream: node scripts/reflect-changes.js solana-core then refresh affected pages.
Validator roles, ledger surfaces, and where transactions enter the pipeline. (Three reading modules — contribution work starts in Level 2.)
Roles across gossip, TPU, TVU, and blockstore—how the node is organized.
Ingress path: stages from fetch through banking and where gossip fits.
Roots, forks, and the surfaces code uses to read blocks and state.
Parallel accounts model and BPF boundary, then a guided first contribution—mirrors Hyperscale Level 2 (two concepts + one OSS slot).
Order: read accounts → read loaders/syscalls → complete the contribution module when ready.
Account model, rent exempt threshold, and scheduler-facing parallelism.
Loader crates, program deployment, and how the runtime dispatches syscalls.
Scoped mergeable slice: docs or tests with clear review surface.
Build and observe like Hyperscale Level 3 orientation: operator/debug skills, then banking/replay literacy, then a scoped contribution with issue hygiene.
Purple module last is intentional—you ship after you can build and read the hot path.
Build from source, run localnet, and follow logs across subsystems.
How banking threads consume packets and how replay confirms ledger state.
Pick a labeled issue or tiny bench—ship with clean repro and tests.