Roadmaps / Linux / Who this is for
Read first
Who this is for
Hours by path (core only; the work version is extra).
| You are | Start at | Skip | Your assignments | Your milestone scope | Aim for |
|---|---|---|---|---|---|
| Fresher (under a year of work; one language, Git, a terminal) | Week 0, then Week 1 | Nothing | Practice projects 1 and 2 in each phase, checkpoints 1 to 4 | Core scope: one service, four seeded faults diagnosed | 2 Sound on every dimension |
| Senior (5 to 8 years, owns services, on call) | Week 3 (skim Weeks 1 and 2; use the starter service below) | Week 9 only if you can already read a packet capture and work out the ephemeral-port limit (Fault 4 needs it) | All practice projects, checkpoints 1 to 4 | Adds the two extra faults, a restore drill and ADRs 5 and 6; optionally Checkpoint W (24 hours) as promotion evidence | 3 Proven on every dimension |
| Staff (10+ years, sets direction across teams) | The fast lane (10 hours, below), then Phase 4 | Phases 1 to 3 weekly tables (the fast lane's gap test tells you which to read) | Checkpoint 4, Checkpoint W and the host-standards memo | Adds a host-standards memo with numbers (8 hours, below) | 3 on all; 4 only through the work version |
| Principal (15+ years, sets strategy) | The milestone section, the fast lane, Phase 4, then Interview topics | The weekly tables of Phases 1 to 3 | Checkpoint W on a real service, plus a review checklist others reuse | Adds a review checklist (6 hours) and a decision memo (8 hours), below | 3 on all; 4 only through the work version |
| Path | Parts | Hours | Weeks at your pace |
|---|---|---|---|
| Fresher | Week 0 (optional, 8), Weeks 1 to 16 | 128 | 16 at 8 hours; 26 at 5; 32 at 4 |
| Senior | Skim Weeks 1 and 2 (4), Weeks 3 to 16 (112), extra faults and restore drill (12) | 128 | 22 at 6 hours; 26 at 5 |
| Staff | Fast lane (10), Phase 4 (32), host-standards memo (8) | 50, plus 24 for Checkpoint W (74) | 19 at 4 hours; 25 at 3 |
| Principal | Fast lane (10), Phase 4 (32), review checklist (6), decision memo (8) | 56, plus 24 for Checkpoint W (80), then a 90-day wait before the recheck | 19 at 3 hours; 27 with Checkpoint W, plus 13 calendar weeks for the recheck |
The fast lane (10 hours) is for Staff and Principal readers who run Linux in production but have not measured it. Use the starter service and provisioning notes in the milestone section (Fault 4 also needs the Week 11 namespace lab, so add about 2 hours), or your own service if it has an HTTP health check and a state directory. The kit: a 10-line unit file with Restart=on-failure, RestartSec=1 and MemoryMax; and a preflight script that checks the kernel has the netem queue discipline (modprobe sch_netem) and cgroup v2 (stat -fc %T /sys/fs/cgroup prints cgroup2fs). Then do the gap test: for each of the six faults in the fault table, write down the first command you would run and the number you expect, run the fault, and mark where you were wrong. Read the Phase 1 to 3 weeks that cover your wrong answers, then go to Phase 4. Checkpoint 4 then runs on your service with the fault table's injection methods.
Prerequisites for a fresher. If one is missing, do Week 0 first.
- You can use a terminal: move around, edit files, pipe commands, read a manual page.
- You can use Git (commit, branch, diff).
- You can write a small HTTP service in one language (a few endpoints, a database call).
- You know what an IP address, a port and a DNS name are.
- You can install a virtual machine from an ISO image.
This roadmap covers general Linux on a server. For Linux on boards and images you build yourself, take the Embedded Linux Engineer roadmap after this one. For the machine itself (CPU, caches, memory ordering), take Engineering Fundamentals.