Forward Deployed Engineer: The Complete Course
Engineering done inside someone else's organisation, on their data, against their constraints, next to the people whose problem it is.
Half the job is code. The other half is working out what to build and getting an organisation to actually use it. This course teaches both, through one customer engagement that runs for twelve weeks, from the first day on site to the check one month after you leave: Larkfield Systems, your employer, deploying an AI claims assistant at Harbourline Freight, a regional carrier. Both companies are fictional; every lesson also opens with a real, cited case.
Who this course is for
You can already ship software, or you are learning to. You want to do it for customers, in their environment, and be the person who makes a product work in the real world.
| You are... | Start at |
|---|---|
| New to building with language models | The AI Engineering course Modules 0 and 7 first, then Module 1 here |
| A software engineer moving into a customer-facing role | Module 1 as a quick check, then Module 2, Lesson 1 and Module 3 |
| An AI engineer who has never deployed at a customer | Module 2, Lesson 1, then Module 3 |
| A solutions or sales engineer who wants to build more | Modules 1 and 2, then Module 4 |
| Already an FDE | The modules where your last engagement hurt |
Every module's README lists what you need to know before starting it, so you can check whether it is safe to skip the one before.
What you need
- A Google account, to run code in Google Colab. Nothing to install.
- One free API key, from Groq or Google Gemini, to use in Colab. Or no key at all, if you run a model on your own computer with Ollama and work in a code editor.
- For a few lessons (a React interface, Docker, a VPN), your own computer. Those lessons say so.
Setup takes about ten minutes: course-guide/02-setup-colab-and-llm-providers.md.
The modules
| # | Module | What you add to the engagement | Needs first |
|---|---|---|---|
| 1 | Engineering Floor | A project skeleton with tests, CI, Docker Compose and safe secrets | Some programming in any language |
| 2 | Speed of Delivery | A cross-system prototype in days, labelled honestly as a prototype | Module 1 |
| 3 | Discovery and Problem Framing | The process map, problem statement, baseline and target | Module 2, Lesson 1 |
| 4 | Their Data | A loader for the TMS export, matched customers, redacted personal data | Module 1's repository; Module 2, Lesson 1 |
| 5 | AI and LLM Application Work | The claims triage assistant, with citations, a human approving, and an eval set | Module 4; LLM basics from AI Engineering Modules 7 to 9 |
| 6 | Enterprise Integration | Single sign-on, roles, service accounts, a release through their process | Module 1, Lessons 2 and 4; Module 4, Lesson 1 |
| 7 | Security and Compliance | A threat model, an audit log, and answers to the security questionnaire | Module 1, Lesson 4; Module 4, Lesson 3; Module 5, Lesson 1; Module 6, Lesson 1 |
| 8 | Domain Immersion | A freight glossary, a schema that fits their business, why the last attempt failed | Module 3, Lesson 1; Modules 4 and 5 |
| 9 | Communication | Status updates, a design doc, demo scripts, a difficult email | Module 3, Lesson 1; Module 5 |
| 10 | Stakeholders and Politics | A stakeholder map and a plan for when your champion leaves | Module 8, Lesson 1; Module 9 |
| 11 | Delivery Management | A risk register, decision log, training, runbook and handover | Modules 3, 9 and 10 |
| 12 | Commercial Awareness | Proof-of-concept exit criteria, an estimate, the value delivered in their numbers | Module 3; Module 5, Lesson 1; Module 10, Lesson 1; Module 11, Lesson 2 |
| 13 | Product Feedback Loop | Product feedback engineering can act on, and one reusable component | Modules 3, 4 and 5 |
| 14 | Operating in the Field | A working plan for site and VPN days, and boundaries that protect you | Module 4, Lesson 2; Module 6, Lesson 2; Module 11 |
| 15 | Professional Practice | A post-mortem of your own engagement and a pattern library | The story of Modules 3 to 14; Module 9; Module 10, Lesson 2; Module 11, Lesson 2 |
How each lesson works
Every lesson is a walk through one Harbourline example, one picture at a time:
- The situation: a specific moment in the engagement.
- A real case, first: something that really happened, from a published source, cited.
- The idea: fifteen to thirty figures, one idea each, one running example.
- Run it, or do it: code with a walkthrough of every line, or a template with a worked example.
- Break it: a failure you cause on purpose, or a weak example you diagnose.
- In production: what goes wrong with real customers, and what it costs.
- Check yourself: questions, one exercise, key terms and sources.
Each module ends with a lab (a guided piece of the engagement) and an incident file (a real, published failure replayed at Harbourline and fixed). More detail: course-guide/01-how-this-course-works.md.
The tools you will use
Python throughout. LangChain for calling language models and LangGraph for workflows with a person approving, so the same code works with Groq, Gemini or Ollama. FastAPI, SQLite, pandas, Chroma or FAISS, and Docker where a lesson needs them. Everything is open source apart from the hosted model services.
Course guide
| File | Read it when |
|---|---|
| 01-how-this-course-works.md | Before your first lesson |
| 02-setup-colab-and-llm-providers.md | Before you run any code |
| 03-the-engagement.md | To meet the customer, the people, the dates and the problem |
| 06-shared-data.md | To see the Harbourline data every module reuses |
| 04-style-guide.md | Only if you are writing or editing lessons |
| 05-image-style-guide.md | Only if you are drawing the course figures |
The one measure
One month after you leave, is it still being used, without you? Almost everything in this course exists to move that number.