Cold Start
Forward deployed work is judged on how fast you ship inside a domain you did not know last month. So I measured it. Every row below is the time from my first commit in a new domain to the first production release, recorded as I worked rather than reconstructed afterwards. One row is measured between different endpoints and says which. Where a start cannot be measured, the row says so.
| Domain | Organisation | Prior Experience | To Production | Scale Reached |
|---|---|---|---|---|
| Procurement data capture | tendr | None | 57 days Nov 25 - Jan 26 | 180 commits, 9 releases |
| Cloud infrastructure | tendr | EVERYANGLE | 4 days Jan 26 | 177 commits, 39 releases |
| Procurement SaaS | tendr | None | 100 days Nov 25 - Mar 26 | 639 commits, 26 releases |
| AI inference engine | tendr | None | 9 days Mar 26 | 268 commits, 19 releases in 43 days |
| Facilities compliance | Handee | None | 24 days May - Jun 26 | 848 commits, 34 releases, 552 test files |
| Fire personnel crewing | Slipway | Retained firefighter | 13 days 23 Aug - 5 Sep 26 | 121 commits, 48 releases, 80 test files |
| Retail computer vision | EVERYANGLE | None | 6-year tenure | Approximate: ~120 repositories, 25+ microservices, multi-cloud |
| EU-funded education | FIP | None | Not measurable | Four-week contract, Aug to Sep 26. Every release tag is the client's |
Reading the record
Four of the six measured starts shipped within 24 days, and none took longer than 100. Four of the six domains I had not worked in. Two rows are marked prior-experience because the same day count on familiar ground is not the same claim: the tendr infrastructure row, and Slipway, which is built for a service I serve in myself. Two of the four sub-24-day starts were cold. The two unmeasurable rows stay on the record because removing them would make the record look better than it is. How the days were spent, and the harness the work runs on, is written up under Method.