Capabilities guide
The capabilities behind forward deployed engineering
Practise the work behind each skill: connect an API, investigate incorrect data, test failure cases, deploy a change, and help someone operate it. When AI is involved, evaluate its outputs and failure modes. Explain your decisions in terms the customer can act on. Each skill below lists the failure mode behind it and where to practise it.
This guide describes human capabilities. The FDEOps AI skill pack is a separate set of instructions for a coding agent. Installing it is not evidence of proficiency.
Technical skills
- Integration work
Read an unfamiliar API, authenticate against it, handle its errors and rate limits, and get data moving between two systems that were not designed to meet.
Failure mode: a happy path that breaks the first time the other side is slow.
- Diagnosing wrong data
Trace one record through both systems, find why the totals differ, and separate what you proved from what you assumed.
Failure mode: trusting a column name instead of tracing an example.
- Testing failure cases
Write tests for relevant failures such as cancellation, duplicate events, missing keys and late files. Rerun the checks affected by a change, then complete the repository's required checks.
Failure mode: demonstrating the cases you already know pass.
- Deploying and monitoring
Use the team's permitted release process, verify the deployed behavior and prepare recovery for the effects the change can leave behind. Monitor the relevant operating cycle.
Failure mode: it ran on your laptop and nobody watched the first real cycle.
- Evaluating AI
Choose representative tasks and checks before relying on a model. Inspect failures, track relevant quality, latency and cost, and distinguish test performance from production evidence.
Failure mode: a convincing demo with no representative checks or explicit action limits.
- Handing over
Leave a runbook, a named operator, a recovery step they have rehearsed, and a list of what is still unresolved.
Failure mode: the only person who can run it is you.
Client skills
- Discovery
Observe the work, ask for a real example and check what the request assumes. Reuse answers already established.
Failure mode: treating a proposed solution as proof of the underlying need.
- Executive communication
State the decision, the tradeoff, and the ask in three sentences a VP can forward.
Failure mode: a status update instead of a decision.
- Scope discipline
Agree what the next change must demonstrate, what it excludes and how new requests will be decided.
Failure mode: leaving scope changes implicit until they cause rework.
- Stakeholder mapping
Understand who uses, funds, approves and operates the result. Confirm authority rather than inferring it from a title.
Failure mode: getting agreement from someone who cannot make the relevant decision.
- Domain learning
Learn enough vocabulary and workflow context to ask useful questions. Validate interpretations with domain specialists and recognise the limits of your expertise.
Failure mode: sounding confident about a domain assumption you have not checked.
- Memory discipline
Log material decisions and their reasoning so later reviews do not reopen settled ground without new evidence.
Failure mode: context that lives only in your head.
A practice rubric for increasing responsibility
This is an editorial self-assessment, not an employer's leveling framework. A person may be strong in one area and still need support in another.
| Dimension | With guidance | Independently in an agreed scope | Helping a wider team |
|---|---|---|---|
| Scope | Clarifies the assigned change | Negotiates a bounded outcome | Helps teams resolve competing needs |
| Discovery | Collects and checks examples | Tests assumptions and recommends a next step | Improves how teams learn from customers |
| Delivery | Implements and verifies with review | Handles relevant exceptions and recovery | Removes recurring delivery obstacles |
| Communication | Explains progress and blockers | Makes decisions and trade-offs clear | Keeps cross-team decisions understandable |
| Continuity | Records useful evidence | Maintains current decisions and ownership | Helps others reuse supported learning |
Keep decisions and evidence usable
Record consequential decisions, their sources and what remains uncertain. Before resuming work, check that the summary still matches the current situation. Keep confidential material within the permissions and systems agreed for the work.
Questions to test your evidence
Choose what to practise next
Pick one capability you cannot yet demonstrate. Use a bounded project, define an observable result and ask someone qualified to review your work. The learning roadmap provides exercises, milestones and resources for each capability. The transition guide helps turn that evidence into an application.
FAQ
What technical skills do Forward Deployed Engineers need?
Integrating an API you did not design, querying unfamiliar schemas, diagnosing data that disagrees between systems, writing tests for the failure cases, deploying through the customer's release process with a rollback, monitoring the first real cycles, and handing the result over. For AI work, add representative evaluation and limits on what the system may do.
Do you need to be a strong coder to be an FDE?
You need the engineering competence required by the role. Be able to explain, test and debug the systems you build. Framework breadth does not substitute for correctness or maintainability.
What non-technical skills matter most for an FDE?
Discovery, clear communication, scope decisions and collaboration help connect engineering to useful work. Keep important decisions and their evidence understandable to other people.
How do I build FDE skills if I am a product engineer today?
Identify a gap in the role you want, then find supported work that exercises it. Customer observation, an integration or an operational handoff can provide useful practice. Record your contribution and seek specific feedback.
Use FDEOps with your coding agent
FDEOps provides reusable instructions for tasks such as discovery, implementation, review and handoff. Use a task on its own, or use the fde coordinator for an ongoing engagement with a local customer record. You remain responsible for decisions, permissions and the evidence behind your claims.