Forward-Deployed Engineering02
One workflow,from prototype to production,in six weeks.
One named senior engineer joins your team, builds one workflow against your real records and cuts it over to real traffic. The sprint ends with a signed runbook and a team that can run it.
- day 01Scope signed
- day 07First pull request
- day 14Golden set agreed
- day 35Cut over
- day 42Runbook signed
- day 01 · scope signed, access requested
- day 07 · first pull request merged
- day 14 · golden set agreed with operations
- Duration
- Six weeks
- Team
- One named senior engineer
- First pull request
- Day seven
- Ends with
- A signed runbook
FDE / 01Fit
Who it is for
The sprint is for you when
- There is one workflow you want in production and a date it matters by
- A prototype or a clear idea of the workflow already exists
- Real access to real systems and records can be granted in the first week
- Someone on your side owns the outcome and can clear a blocker in a day
- Your team can take the handover: at least one engineer who will run it
It is not when
- The workflow has not been chosen yet; the Readiness & Scoping Sprint comes first
- Access will take a quarter to arrange
- The problem is a research problem no model yet solves
- You need capacity on a backlog rather than one outcome
- There is nobody to hand the system to at the end
FDE / 02Week by week
W0 – W6
How the six weeks run.
Six phases, each ending in something you can hold. The diamonds are the gates: scope signed, cut over, handed over.
Week two is the last week anything runs on sample data. From week three, every run is against your records.
FDE / 03What exists at the end
6 ITEMS
Six things, none of them a slide.
The sprint is complete when every item on this sheet is ticked by your side, not ours.
- The workflow in production
Running on real traffic under your account, with rollback armed and monitoring live.
- The repository
Every commit in your repository, reviewed by your reviewers, intellectual property assigned before week one.
- The evaluation suite
The golden set, the scoring harness and the regression run, wired into your pipeline.
- The runbook
Deploy, rotate credentials, retrain, roll back, escalate; executed by your engineers in the handover dry run.
- The decision record
Each architectural decision, the options considered and why one was taken, dated.
- The owner who has run it
One named engineer on your side who has released, rolled back and retrained the system without us on the keyboard.
FDE / 04How the sprint is run
Four rules.
They keep six weeks from becoming twelve. Each is agreed at scoping and holds until the runbook is signed.
The scope does not grow
One workflow, one metric, one owner, on one page. Anything discovered along the way is written down for the next sprint, not added to this one.
Access is the first blocker, so it is cleared first
Repository, systems and records are requested on day one. If access is not real by the end of week one, the sprint pauses and the clock with it.
Nothing ships that has not been scored
The golden set is agreed with the people who do the work by the end of week two. Every candidate is scored against it before it reaches a user, and the score is what earns the cut-over.
Your team does the handover, we watch
In the last week your engineers perform a release, a rollback and a retrain with our engineer in the room but not on the keyboard. If that does not go cleanly, the sprint is not finished, whatever the calendar says.
FDE / 05Specified
10 FIELDS
The sprint, specified.
The terms as they stand before the first call. Scope moves the detail; it does not move the shape.
- Duration
- Six weeks from the day the scope is signed
- Start
- A scoping call, then a one-page scope agreed with the owner of the outcome
- Engineer
- One named senior engineer, full time, met before you sign
- Cadence
- Your standups daily; a written report to the owner weekly; a demonstration on real records at each gate
- Access needed
- Repository, the system of record, real records and a person who can approve a pull request
- Where the work happens
- In your repository and your cloud account; on site for scoping and embedding where the work needs it
- Evaluation
- A golden set agreed by the end of week two; every release scored against it
- Cut over
- Week five, on real traffic, with rollback armed and monitoring live before the switch
- Handover
- Week six: runbook, decision record and a dry run performed by your engineers
- If it slips
- For reasons on our side, we stay until the gate is met at no further fee; for access or decisions on yours, the clock pauses and we write down where it stands
FDE / 06Proof
2 ROWS
Case studies
Illustrative engagements written end to end, chapter by chapter.
FDE / 07Questions
6 QUESTIONS
Asked about the sprint.
Do we need a prototype before the sprint starts?
No. A clear idea of the workflow is enough; week two builds the prototype on your stack. If the workflow itself has not been chosen, the Readiness & Scoping Sprint comes first and ends with a scope this sprint can start from.
What does the engineer need from us in week one?
Repository access, access to the system of record and real records, a seat in your standups and one person who can approve a pull request. Access is what causes slippage, so it is requested on day one.
Can the scope change during the sprint?
The workflow, the metric and the owner do not change. What is learned along the way is written down for the next sprint. That is how six weeks stays six weeks.
What does cut-over involve?
Real traffic at real volume, with rollback armed and monitoring live before the switch is made. The golden-set score is the argument for going ahead; if it is not passing, the cut-over waits.
What happens after week six?
Your team runs the system; the runbook and the decision record are theirs. Our engineer stays on call for what was built for an agreed period, then steps back. Managed Operations exists if you would rather we keep running it.
How is the sprint priced?
As a fixed fee, agreed before week one against the written exit gate. There is no hourly billing and no invoice for a gate that was not met.
FDE / ENDStart
Six weeks from the first call,one workflow runs in production.
Book a scoping call; you leave it knowing whether the workflow fits and what it would take.