Chapter 00: Before Week 1

Ninety Days From AIML Graduate To Forward Deployed Engineer

Not another framework tutorial. A sequenced plan (build, break, evaluate, deploy) that ends with three working systems, close to a hundred problems you've actually solved, and an interview process you've rehearsed before anyone else asks you a question.

Twelve weeks, three months, one sequence. Each week has a short list of concrete tasks and exactly one deliverable. The deliverable is the point: it proves you did the week, and it eventually becomes your portfolio.

Start with Week 1 from the sidebar, or jump straight to the hundred problems using the shortcuts on the right. Either way, don't start Week 2 with an unfinished Week 1. A half-built thing teaches you less than a small finished one.

01

Zero to Forward Deployed Engineer, 90-Day Roadmap

→ $170K to $450K+: reported total-comp range across levels for forward-deployed / agentic AI engineering roles, US market, 2026.

→ 800%: growth in forward-deployed AI job postings, Jan to Sep 2025, as companies learned demos don't close deployments, delivery does.

→ 88%: share of agentic AI pilots reported to stall before production. The exact gap this role exists to close.

→ 60% to 25%: how agent success drops once a task must succeed eight times in a row, per Sierra's tau-bench. The reliability cliff you're training to fix.

One assumption, stated up front: this plan treats you as someone who can read and write basic Python from an AIML degree, but has never shipped a real backend or touched an agent framework. If that's wrong in either direction, the sequence still holds; you'll just move through some weeks faster than others.

02