MoJ Justice AI Unit recruits five forward-deployed engineers at £71-£85k
TL;DR: The Ministry of Justice has opened applications for five “forward-deployed engineer” posts inside its Justice AI Unit — the in-house team building AI tools for courts, prisons and probation services. Salaries run from £71,381 to £85,257. The advert promises end-to-end product ownership “from problem discovery to product deployment and scale”, with applications closing at 23:55 on 28 September.
The job description leans toward agentic engineering rather than data-science modelling: candidates will work “directly in the field” with frontline staff in courts, prisons and probation services, then “develop high-impact AI products, models and agents that will be widely used”. The unit’s stated remit is to “build and ship products quickly across the Ministry of Justice”.
What the move signals
Five engineers does not sound like much, but the role profile and salary band tell a clearer story. £85k tops out below market rate for senior AI engineers in London — typically £120-200k base in private sector — but matches Civil Service ceilings for unbadged technical posts. The MoJ is positioning the work as mission-driven, with the advert explicitly soliciting candidates who can “work in ambiguity” and “learn quickly on the job”. That framing is closer to a startup than a traditional Civil Service hire.
The posts also hint at a structural shift. Earlier MoJ AI work was outsourced to consultancies and prime contractors. A forward-deployed in-house team — the term itself borrowed from frontier-AI vendor Anthropic and Palantir’s deployment-engineer model — implies the department wants vendor-style velocity inside the perimeter, on its own data, with ministers accountable. PublicTechnology notes earlier MoJ commitments to “embed” AI department-wide, but this is the first time an explicit engineering squad has been advertised.
Looking forward
Expect this to be the leading edge of similar units across UK central government. The Justice AI Unit will be watched closely as a test case for whether civil-service salary bands can attract retainable AI engineering talent against fintech and frontier labs, and whether forward-deployed in-house engineering produces measurably faster deployment than the prevailing consultancy model. For UK GovTech vendors and SaaS suppliers, the message is straightforward: government departments are insourcing the work that used to be project-managed out. Suppliers will need to compete on tooling, data access and workflow integration rather than on delivery hours.