You’re approving the AI roadmap.
You don’t need to write it.
You do need to read it.
The same course that trains FDEs gives finance principals the technical literacy to evaluate, hire, and deploy them — without faking expertise you don’t have. Same 27 modules. Same capstone. Same 33-question interview bank. Different engagement pattern.
Mid-level engineer (2–8 YOE) pursuing an FDE role at Palantir, Anthropic, OpenAI, Ramp, Scale, or Databricks.
Licensed finance professional — CPA, FP&A, treasury, corp dev — using domain depth as the wedge into an FDE role.
CFO, controller, partner, finance director — approving the AI roadmap, hiring the people who build it.
Not an executive overview. Not an AI-for-finance-leaders repackage. Not MBA-style.
Most executive-education AI courses are theater. They’re sized for a busy principal’s calendar — eight hours, certified by a brand-name school, comfortable. They’re not sized for competence. After completion, the principal still cannot tell whether a candidate is demonstrating the moves or reciting them. The certificate is the deliverable.
This course is the actual course an FDE candidate reads — the same 27 modules, the same capstone rubric, the same interview bank. The difference is how you engage with it. You don’t ship the project; you read the rubric. You don’t write the eval; you ask whether the eval is causally honest. The exec-ed product is sized for completion. The material is sized for competence.
Three falsifiable outcomes.
Each is testable in a real meeting, not in a self-rated quiz. Each maps to a specific phase of the course.
The Phase 4 capstone scores submissions against ground truth using a causal-uplift metric — explicit because most AI projects ship with vanity metrics that reward contacting sure-thing payers. After reading the capstone rubric, you can sit in an architecture review and tell whether the success metric your team proposed is causally honest or whether it would have looked good even with no model in the loop.
The Phase 5 interview bank teaches what a strong answer demonstrates — the shape of the thinking, the named moves — not the answer itself. After reading it, you can run three or four questions from the bank in a real Applied-AI interview and know whether the candidate is demonstrating the moves or reciting a framework they memorized.
Phase 3 treats evals as the load-bearing discipline. After reading modules 2 through 4, you can sit in an AI/ML proposal review and ask the questions vendors don’t want to answer: what does the eval set look like, what’s the false-positive cost, how is performance going to be measured against the baseline you would have had with no model.
Same 27 modules. Different engagement.
- P-01
Foundations
Skim. You won’t write the Python; you need to recognize the dev workflow your engineers should follow and the failure modes that distinguish a shipped artifact from a notebook.
- P-02
Technical Specialization
Read for vocabulary. When the team says “we need a vector DB” or “we should fine-tune,” you should know what those mean and whether the answer is sized to the problem.
- P-03
AI / ML Specialization
Read modules 2, 3, and 4 carefully. Skip the project. Then demand any AI/ML proposal you sign off on submits a project rubric that holds up against the eval criteria here.
- P-04
Client Engagement and Product
This phase is about you. The FDE candidates you’ll interview are being taught how to read your team, your data, and your political constraints. Read what they’re being taught.
- P-05
Portfolio and Job Readiness
The interview bank is the rosetta stone. Walk every Applied-AI candidate through three or four questions from it in real interviews. You’ll catch the ones who recite vs. the ones who demonstrate.
Not executive education. Not theater.
Most AI training built for finance leaders fails the same way. It abstracts the engineering to make the principal feel comfortable. It teaches metaphors instead of machinery. The result is a finance leader who can hold a conversation at a conference and still cannot tell whether the rubric their team submitted is real.
The course doesn’t do that. It gives you the same primary sources, the same code, the same rubrics that the engineers reading their own copy are working through. You don’t fake the engineering. You read it. The competence you walk out with is the competence to evaluate work, not the competence to produce it — which is what the seat actually requires.
Take the team’s training seriously.
The same sign-in as your engineers. The same modules they read. Read them differently.
Start the course