context_window.md — about / curriculum source
Every lesson starts as a ticket a developer could meet on Monday morning.
About and How We Teach
How we keep AI coding lessons honest
AI Code Pathway Academy is a small training group in Cambridge, Massachusetts, teaching AI-assisted software development as developers practice it in 2026 — resolving in real time, ticket by ticket, with review built in.
Watch a diff rewrite itself, then learn why — who writes it, how it stays current, and how work is reviewed.
Who stands behind the material
Who builds the lessons
Curriculum is written by practicing developers who teach part-time, with an editor who checks every exercise against the current syllabus. Lesson authors work from the same repos learners use, so instructions match what is on screen. Bios list current teaching roles and recent course work, not honorary titles. If an author has not run the exercise, it does not ship.
author runs the ticket
reviewed
Requirededitor checks syllabus
each term
Requiredbios show real work
current
OpenWhy we stay narrow
We teach three tools plus the agentic methodology that connects them: Learn Claude Code, Learn Cursor AI, and Learn OpenAI Codex. AI coding training stays deep when the list stays short, so we revise deeply each term instead of skimming many products. Related assistants appear where comparison helps secure or enterprise context, but they are not separate tracks. Depth over catalog size is the point.
To Learn AI-powered development well, you need repeated practice in one workflow — plan, run tests, read the diff — not a tour of dashboards. That is why our AI coding courses keep the same repos, the same review habits, and the same bar across every track.
How often material changes
Each term we retire exercises that no longer match current tool behavior and rewrite prompts that produce noisy diffs. Updates draw on applied training trends visible in 2026 catalogs, applied to our own repos and checks. The changelog inside the syllabus shows what changed and why.
Learners always see the current version before enrolling. When a Claude Code course lesson or a Cursor AI course exercise is replaced, the old ticket is archived with its reason, and the new one ships with fresh reviewer notes.
In the room
How class time actually runs
Evenings in Cambridge mix short demos with long practice blocks. Online cohorts follow the same plan with smaller review rooms.
evening_session.plan
Two instructors attend each session: one leads, one reads diffs with learners. That second pair of eyes is how review habits get coached in real time.
- short demo, then long practice block
- OpenAI Codex course tickets worked in pairs
- reviewer notes kept with your work
online_cohort.plan
Online cohorts run the same tickets with smaller review rooms, so every learner gets time on the diff. You bring your current stack; we bring the structure.
- same syllabus, smaller rooms
- plans written before code runs
- tests run, diffs read aloud
The exchange turns mid-page: one room leads, the other reads — and the habit sticks.
Evaluation
How we evaluate without hype
No inflated promises. Completion means work you can show.
completion record — kept by the learner
Completion reflects attended sessions, submitted ticket work, and demonstrated review habits: plans written, tests run, diffs read. We do not predict hiring, exam, or salary results, and we do not issue accredited degrees.
What you keep after the course is concrete: your worked tickets and reviewer notes, ready to discuss with your team.
Who we serve best
Working engineers who want agentic workflows inside their current process. Teams that need shared review standards for AI-assisted pull requests. Career switchers with code fluency who want structured practice before using these tools at work. Beginners without code background start with prep, not with the main tracks.
Where to find us
Our classroom is at 77 Massachusetts Ave, Cambridge, MA 02139, USA. Phone +1-617-253-1000. The nilxoxron you are reading, nilxoxron.sbs, is our only site address; course mail comes from the same name. One business, one address, one contact path.
Monday to Friday, 9:00 AM – 6:00 PM · [email protected]
What to do next
Read, request, then ask us anything unclear.
Read the track pages, request a syllabus, then ask us anything unclear. We answer with specifics about schedule, prep, and fit. If another option suits you better, we will say so.
Start with the three tracks — Learn Claude Code, Learn Cursor AI, Learn OpenAI Codex — then compare them against the agentic guide. Bring questions about your stack and we will point to the right first ticket.
failing test → passing diff · one session
Prefer a person? Contact the teaching team — or return Home.