AI Engineer
Keter AI is a senior-led AI consulting and engineering boutique that takes organisations from AI ambition to working systems, with strategy, product engineering, sovereign architecture and governance in one accountable team.
As an AI Engineer you build the LLM and agent systems our clients run in production: retrieval over their own documents, agents that act through their tools, and the evaluation, guardrails and integration work that makes those systems safe to rely on. Some engagements run in public cloud; others run in EU-cloud, private or air-gapped environments where data cannot leave. You work in Python and TypeScript, alongside the client's own engineers, from first prototype to handover.
What we look for
Four or more years of professional software engineering, mainly in Python with working TypeScript, including LLM-based systems you have shipped to production and kept running.
Retrieval done properly: document parsing that copes with tables and scans, hybrid search and reranking, permission-aware retrieval, and answers that cite their sources.
Evaluation as a habit: you build test sets from real user questions, measure retrieval and answer quality, track regressions and gate releases on the results.
Agents and tool use: multi-step workflows with tool calls, approvals, escalation and a human in the loop, and the judgement to choose a deterministic workflow when it is the better answer.
Guardrails and security: input and output controls, defences against prompt injection, careful handling of personal data, and logging, tracing and an audit trail that someone can actually read.
Enterprise integration: APIs, identity and access, document stores, relational databases such as PostgreSQL, queues, Docker, CI/CD and at least one major cloud.
Sovereign and on-prem deployment, or a serious interest in it: serving open-weight models on private infrastructure, sizing hardware, and weighing hosted against self-hosted models on cost, latency and data residency.
Consulting habits: you can explain a trade-off to a CTO and to a compliance lead, write a clear design note and work in fluent English. Polish, and working knowledge of GDPR and the EU AI Act, are an advantage.
What we offer
EUR 5,000-7,800 per month (about USD 5,700-8,700), depending on seniority, on a long-term, full-time B2B contract.
Fully remote work within European time zones. Occasional client visits are planned in advance, and travel costs are covered.
A senior team: you work directly with the people who scope each engagement, with no layers between you and the client's engineers.
Real client work across the full arc of strategy, architecture, delivery and governance, from first prototype to systems in production.
Varied problems, from regulated sectors and sovereign or on-prem deployments to AI Act readiness, rather than one product's backlog.
An annual learning budget for courses, conferences, certifications and model or compute experiments.
The selection process
Introductory call (30 minutes). Your work so far, what you want to do next, and how we run client engagements.
Technical conversation (90 minutes) with our engineers. You walk us through a system you built, then we design one together from a realistic client brief, covering retrieval, evaluation, guardrails and deployment.
Final conversation. We agree the first engagement, the working arrangement and the terms. We reply after every step.
Before you apply
Send a short note through our contact page with links to work you have delivered: a repository, a write-up, a demo or a description of a system in production. Tell us what you built, what you would change now, and which time zone you work in.