R-05 · RACD
RESEARCHFounding Research Engineer, RACD
The bar is near-zero false positives. You cannot hit a bar you cannot measure, and measuring this one honestly is its own engineering problem.
FULL-TIME · FOUNDING · EVALUATION INFRASTRUCTURE · REMOTE · OVERLAPPING HOURS
OPENINGS
1
One. When it is filled, this page comes down.
BASE SALARY
£215,000 to £355,000
The band, published. The offer is the same whether or not you push for more.
EQUITY
0.45% to 0.95% of RACD
Founding equity in the company you would actually work on.
01
The problem this role exists for
RACD's entire claim is an alarm an engineer would act on at 03:00 without checking first. Proving that requires replaying real estates against candidate models, generating failures that are realistic rather than convenient, and producing a number nobody in the room is able to argue with. Most research groups discover far too late that their evaluation had been flattering them.
You would build the machinery all of the research runs on: the replay harness, the simulated estates, the fault injection, the benchmark suite, and the experiment tracking that makes a result reproducible by somebody who was not there. This role decides how fast everyone else can move, and it is treated here as research infrastructure rather than as tooling.
More on the company itself: RACD →
02
Your first week, and a normal Tuesday
Your first week is spent reproducing somebody else's result from three months ago. The exercise is chosen deliberately: it is the fastest way to see exactly where the current setup is dishonest, and your write-up of what you could not reproduce becomes the first version of the roadmap.
A normal Tuesday: extending the harness so a new class of estate can be simulated, a review with a scientist whose experiment needs something the harness cannot yet express, and an unglamorous stretch making the whole thing faster, because an evaluation that takes six hours is an evaluation people quietly skip.
WHAT YOU WOULD BE WORKING IN
Python throughout, something compiled where the replay path is too slow, a columnar store full of real telemetry, containers, and a cluster with genuine GPU budget behind it. Experiment tracking and data versioning tools are your call rather than something you inherit.
We do not screen on tools. Nothing above is a requirement, and a stack list is a poor proxy for whether somebody can do the work. If you have gone deep in something else and would arrive with a view about why half of this is the wrong choice, that is an argument we would like to have.
03
What you would own
Not a ticket queue and not a slice of somebody else’s roadmap. Four things with your name against them in the decision log, from the first fortnight.
01
The harness
Replay of real estates against candidate models, fast enough that people use it.
02
Simulated failure
Faults that are realistic rather than convenient, including the awkward ones.
03
The number
One benchmark result nobody in the room is able to argue with.
04
Reproducibility
Somebody who was not there reruns it and gets the same answer.
04
What we are looking for
MINIMUM
What we would need to see
- ▸You have built research or evaluation infrastructure that other people depended on daily
- ▸Strong Python, and the engineering discipline that research code usually lacks
- ▸You understand why a benchmark flatters its authors, and you design against it
- ▸You can work with researchers without either deferring to them or dismissing them
PREFERRED
What would move you up the pile
- ▸Simulation, fault injection or chaos engineering experience
- ▸Experiment tracking, data versioning and reproducible pipelines at a scale where it mattered
- ▸A background in science or engineering where measurement itself was the hard part
- ▸You have caught a result that was wrong, and you can tell that story
Nothing in the preferred column is a filter, and neither list is applied before a human reads you. No degree is required for anything on this page, and we have hired people with doctorates and people with no degree at all into the same rooms. If you meet most of the minimum and not all of it, write anyway and say in the first paragraph which part you do not meet.
05
Pay, and everything around it
£215,000 to £355,000 base, plus 0.45% to 0.95% of RACD, reviewed upward annually without you having to ask. The band is published here rather than discovered three conversations in, and the offer is the same whether or not you negotiate, because the band is the band.
Everything else is what the whole group gets and is not negotiated per offer: a holiday floor rather than a ceiling, health and dental and vision for your family, sixteen to twenty weeks of parental leave for every parent, a funded setup, monthly paid group meetups somewhere in the world, one required funded trip abroad each year, a protected day a week on the research ventures, and no non-competes.
The full list, the salary band chart and the visa sponsorship map are on the careers page. We sponsor without qualification across seventeen jurisdictions and publish exactly which.
06
Apply for this role
Four steps: a conversation with somebody who would actually work with you, one real problem drawn from work we are genuinely doing, a working session on your solution, and a decision within a week of that session. The problem is paid if it runs long, and you keep the work either way.