Choose a first FDE portfolio project with a bounded client problem

The best first project is small enough to evaluate end to end. A broad autonomous CRM agent creates permissions and integration risk before the learner can demonstrate the core decision.

Score two fictional options

The delivery lab weights owned data, bounded writes, testability, demo value and access risk.

from delivery_cases import project_choice_case

result = project_choice_case()
assert result["selected"] == "invoice-review"
assert result["scores"]["invoice-review"] > result["scores"]["live-crm-autonomy"]
assert result["fictional_client"] is True
assert result["external_access"] is False

Choose a task with a clear current process, synthetic/rights-cleared data, measurable acceptance, useful exception path and reversible proposal. Good examples include document review, cited knowledge retrieval or read-only operational summaries.

Avoid claims of a real client, production savings or deployment unless evidenced. Build normal, unsupported and failure paths; include test data, architecture decision, cost model, threat model and runbook. A smaller complete system demonstrates more engineering than disconnected features.

The FDE for Freshers course culminates in a fictional enterprise engagement with explicit evidence.

Exercise

Score three project ideas with your own weights. Select one, write the non-goals and identify the critical regression that must block the demo.

Continue learning

This article is part of the FDE career entry and client-delivery practice sequence. Use the neighbouring tasks when you need the prerequisite or the next application.

• Review the prerequisite or neighbouring task in FDE readiness for freshers: demonstrate foundations before autonomy.
• Continue with Run a discovery interview for a fictional client.

Reference: GOV.UK service assessment preparation.