Mentorship
The accelerator teaches in a gradual release model. Units are I do. Evidence gates are you do. The Mentor Corps staffs the band in between — we do — which is where veterans stall and where no automated system reaches.
Your mentor is a working engineer: a Vets Who Code graduate currently employed in the field, or a veteran engineer from outside the accelerator. Either way, someone who has already done the thing you are trying to do.
Two tracks
Track 1 — Technical. Pre-gate artifact review, architectural tradeoffs, debugging workflow, git CLI, and the question that matters most: would this survive code review at a real company. Coverage follows the curriculum — React and TypeScript on the frontend, the data layer including pgvector, FastAPI on the backend, and AI Engineering: RAG, agents, production AI APIs.
Track 2 — Translation. Turning mission experience into engineering impact statements, mock technical and behavioral screens, portfolio review, and offer negotiation. It runs light through the accelerator and becomes the main event at graduation.
What it is not
Mentorship is scoped deliberately above syntax. You already have the harness — OpenCode and CodeGemma 2B — for syntax questions, and it answers faster than a human can. Mentor time goes to judgment: the calls a model cannot make for you.
It is also not a networking app and not a coffee chat. It is a standing appointment with a specific engineer who knows what you shipped last week.
What you get
- One 45-minute sync every week, on a schedule you and your mentor set
- One pre-gate artifact read per unit, returned within 48 hours
- Async access in the cohort Discord, within your mentor's stated hours
- Six months of Track 2 support after graduation — mock screens, resume, interview loops, offer negotiation
Free, like everything else we run.
How it runs across 17 weeks
Onboarding, weeks 1–2. Confirm the match, review your weakest units from intake, and agree on when you escalate rather than grind. You bring learning objectives, a working GitHub, and your first gate submission.
Frontend, weeks 3–6. The highest-touch phase. This is where attrition happens, so this is where the mentor leans in hardest — unblocking and confidence, not lectures.
Data layer, weeks 7–10. Correctness review. Bad schema and query decisions fail quietly, and gates alone will not catch them. You explain your modeling choices out loud.
Backend, weeks 11–14. Tradeoff conversations, API contracts, CI/CD, and what production actually looks like. You walk your mentor through the system design of your own work.
AI engineering, weeks 15–17. RAG and agent design judgment, and the difference between production-real and demo-real. Portfolio goes final and the job search opens.
Post-graduation, six months. Track 2 takes over while you apply, interview, and negotiate.
Data structures and algorithms run as a continuous thread across all seventeen weeks, on their own recurring touch, because interviews are shaped differently than the work.
Your mentor never grades your gate
This one is structural, not a courtesy. A veteran will not admit "I don't understand this unit" to the person who decides whether they pass. Grading and mentorship stay in separate hands so the weekly sync is a place you can be honest about what is not landing.
Mentors review against the published rubric for each unit — not personal preference — and are briefed on the failure modes that make mentorship worse than nothing: writing the code for you, drifting into life coaching, and contradicting the curriculum.
Getting a mentor
Mentorship comes with the accelerator. If you are in a cohort, you are matched during onboarding.
If you are an engineer who wants to serve on the other side of this, the commitment is one cohort at four to five hours a month — see Mentor with us.