Forward Deployed Engineer
intermediate-advancedThe customer-facing engineering track: embed with a customer, find the problem worth solving, get their data out and clean, build the integration and the AI workflow on their own ground, prove it with evals on real cases, deploy it inside their walls, and hand over a system their team can run. Eight weeks, seven graded projects, one fictional customer you deploy against from discovery to handover.
Never written code? You can still start here.
Module 0 opens with 18 lessons of programming from nothing — your first program, variables, loops, functions, reading errors, the terminal, Git and arrays — before any of the Forward Deployed Engineer material. No prior coding assumed.
What you'll learn
~102 h guided · about 21 weeks at 5 h/week · +74 h on projects
- Explain why ChatGPT and Claude generate answers rather than looking them up like a search engine
- Describe the forward deployed engineer role in one sentence a customer and an engineer would both accept
- Plan the first five days on site so they produce evidence rather than impressions
- Choose the extraction route for a system from its interface, its owner and the access you can actually get
- Wrap every external system behind an adapter with one interface the deployment owns
- Place the model at one step of the customer's process, with a defined input, output and fallback
- Write a definition of correct for each workflow output that the customer's process owner will sign
- Identify which of the four deployment targets a customer requires from their data, their regulator and their history
Curriculum
0/64 lessonsAI Foundations — optional basics (free)
Week 0 · 20 lessons
Beyond the chatbox: what AI assistants really are
How the model "thinks" — and why it changes how you ask
The capability map: what AI does brilliantly
The limits: hallucination, freshness, and what never to trust it with
Your first real win: turn a 30-minute task into 3
Context is everything: give it what it needs to know
Give it a role and a goal
Constraints and format: define what "good" looks like
Show, don't tell: examples and templates
Iterate like a pro: steer, don't restart
Documents and reports: summarise, extract, and draft better
Spreadsheets and data: Excel formulas, analysis, and charts
Presentations: from blank page to a solid deck
Long, messy input: transcripts, threads, notes, and screenshots
ChatGPT, Claude, Copilot, Gemini: which to reach for, and when
Make smarter decisions: AI as your thinking partner
Chaining steps: multi-part tasks done reliably
Your prompt library: reusable templates that save hours
Staying safe: privacy, accuracy, and company data
Capstone: do a real task end-to-end (graded) → your certificate
Module 0 — Are you ready?
Week 0 · 4 lessons
What a forward deployed engineer actually does (and is not)
What this course assumes: Python, APIs, SQL, git, and one model API
The two muscles: engineering judgement and customer judgement
If this is too soon: where to start instead
Week 1 — Discovery: finding the problem worth deploying against
Week 1 · 5 lessons
Landing on site: the first week with a customer
Discovery interviews that produce specifications
Mapping the customer's systems and data before promising anything
Scoping: the smallest deployment that proves value
Writing the deployment brief: success metric, owner, deadline, risks
Week 2 — The customer's data
Week 2 · 5 lessons
Getting data out: exports, APIs, databases, and the people who guard them
Profiling messy enterprise data before you build on it
Entity resolution: the same customer in four systems
Building a reliable ingestion pipeline (idempotent, resumable, logged)
Data contracts and the handover the customer's team can maintain
Week 3 — Integration engineering
Week 3 · 5 lessons
Integrating with systems you cannot change: CRMs, ERPs, ticketing, legacy
Auth in enterprise land: OAuth, service accounts, SSO, key rotation
Webhooks, polling and queues: getting events out reliably
Rate limits, retries, and the backfill that takes a weekend
Sandbox to production: the promotion path
Week 4 — Building the AI workflow on customer ground
Week 4 · 5 lessons
From demo to workflow: where the model sits in the customer's process
Prompting against the customer's own documents and vocabulary
Retrieval on their corpus: chunking, permissions, freshness
Structured outputs and tool calls into their systems
Human-in-the-loop design: who approves what, and how it is logged
Week 5 — Proving it works: evals on real data
Week 5 · 5 lessons
Defining correct with the customer
Building a golden set from real cases (and getting it signed off)
Eval harness: per-slice scoring, regression gates
Measuring value: baseline, before and after, cost per task
The demo that is actually a measurement
Week 6 — Deploying inside the customer's walls
Week 6 · 5 lessons
Deployment targets: SaaS, VPC, on-prem, air-gapped
Containers, configuration, and secrets across environments
Security review: the questionnaire, the pen test, the data-flow diagram
Observability the customer can read: logs, traces, dashboards
Incident response when you are the vendor
Week 7 — The people side of deployment
Week 7 · 5 lessons
Running the pilot: users, training, feedback loops
Change management: why good tools go unused
Executive communication: status, risk, and the numbers
Stakeholder mapping: champion, blocker, economic buyer
Saying no: scope creep, feature requests, and the product feedback loop
Week 8 — From one deployment to many
Week 8 · 5 lessons
Generalising: what was bespoke, what becomes product
Playbooks and runbooks: making the second deployment 10x faster
Handover: leaving a system the customer's team can run
Commercial awareness: renewals, expansion, and what you are measured on
Your FDE portfolio: writing up a deployment as evidence
What you'll build
7 projects · ~74 h · each graded against a rubric
The Discovery Pack
intermediate · ~6 h
Pythonpandas or csvMarkdownThe Ingestion Pipeline
intermediate · ~10 h
PythonSQLite or PostgrescsvYAMLThe Integration
intermediate · ~10 h
PythonHTTPHMACSQLThe Workflow
advanced · ~12 h
PythonA model APIJSON Schema or pydanticSQLThe Eval
advanced · ~10 h
PythonJSONYAMLMarkdownThe Deployment
advanced · ~12 h
DockerPythonYAMLJSON logsCapstone: The Handover
advanced · ~14 h
MarkdownPythongit
What we track
Everything below is graded by Nova and kept on your record — the thing an employer can open, run and verify.
- Every lesson completed →
Mastery per lesson from the spaced-retrieval cards, and your streak.
- Every exercise, graded →
Nova grades each one and keeps the grade — no self-marking.
- Every project, scored against a rubric →
The score, the feedback and the repo, on your portfolio for an employer to run.
- Your skill report →
The placement assessment maps your level topic by topic and updates as you re-check.
- The certificate →
Issued on completion, verifiable by anyone at /verify, addable to LinkedIn.
Get started
A senior, self-paced track — no placement test. Jump straight into lesson 1, free.
Featured Projects
The Discovery Pack
The Ingestion Pipeline
The Integration
About
The people behind this track
Nova grades every submission; these are the humans who wrote the brief and stand behind it.