R-04 · RACD
RESEARCHFounding Distributed Systems Engineer, RACD
Correlation across an estate is a distributed systems problem wearing a statistics hat. Clocks disagree, sources vanish, and the answer still has to be one alarm.
FULL-TIME · FOUNDING · REMOTE · OVERLAPPING HOURS
OPENINGS
1
One. When it is filled, this page comes down.
BASE SALARY
£230,000 to £390,000
The band, published. The offer is the same whether or not you push for more.
EQUITY
0.50% to 1.05% of RACD
Founding equity in the company you would actually work on.
01
The problem this role exists for
RACD reads from thousands of sources across estates it does not control. They disagree about time, they drop out without warning, they replay old data after an outage, and a correlation engine that assumes any of that away produces confident nonsense. The hard part is not the model. It is building a substrate where the model's inputs mean what they claim to mean.
You would own the runtime the correlation work sits on: how state is partitioned, how late and duplicate data are reconciled, how the system degrades when a third of an estate goes dark, and how any of it recovers without a human deciding what to replay. The product promises one alarm per event, and most of the ways that promise gets broken are yours to prevent.
More on the company itself: RACD →
02
Your first week, and a normal Tuesday
In your first week you are handed the current failure list, which is honest and longer than we would like, and asked to pick the one you think is most load-bearing. Something of yours goes to production inside the fortnight, because reading about this system teaches you considerably less than breaking it does.
A normal Tuesday mixes design and forensics: an hour writing up how you propose to handle a class of out-of-order arrival, then several hours working out why one customer's estate produces a shape nobody predicted. You will pair with the research scientists more than you expect, because half of what looks like a systems bug is a modelling assumption and the other half is the reverse.
WHAT YOU WOULD BE WORKING IN
A compiled language on the hot path, Python where research meets it, a streaming layer, a columnar store, Kubernetes. Rust and Go are both here and we have not tried to pick one. Depth in the ideas matters far more than familiarity with any particular runtime.
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 runtime
How correlation state is partitioned, checkpointed and recovered.
02
Time
Late, duplicated and replayed data reconciled into something a model can trust.
03
Degradation
What happens when a third of an estate stops reporting, stated rather than discovered.
04
The failure list
Currently honest and longer than we would like. Yours to shorten.
04
What we are looking for
MINIMUM
What we would need to see
- ▸You have built or operated a distributed system where correctness under partial failure mattered, and you can describe a specific way it went wrong
- ▸Working knowledge of consensus, replication or stream processing semantics, and where exactly-once claims stop being true
- ▸Comfort in a compiled language, and no strong attachment to which one
- ▸You debug from evidence rather than intuition, and you can show your working
PREFERRED
What would move you up the pile
- ▸Event-time processing, watermarking, or reconciling out-of-order streams at scale
- ▸You have worked on observability or telemetry infrastructure from the inside
- ▸Experience running something whose recovery story was tested rather than assumed
- ▸Interest in the statistical layer above you, because the boundary is where the good bugs are
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
£230,000 to £390,000 base, plus 0.50% to 1.05% 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.