AI Systems Engineer
-
Software Development
Job Description & How to Apply Below
Who we are:
Nimble is a boutique technology consultancy based in the UK North;
Sheffield, Leeds and
Manchester. We partner with ambitious clients across the public and private sectors to design,
build, and optimize digital services that deliver real value. We pride ourselves on our agile
mindset, our creative problem-solving, and our collaborative culture.
Our projects are diverse and different, just like our ways of working and our people. One of our
key focusses is maintaining a strong culture of pragmatic, skilled, egoless people who collaborate to get things done and have lots of fun along the way.
The people we wantWe want curious, tenacious problem solvers who create forward momentum. We value people
with a high learning capacity: people who get to the real problem, learn fast in unfamiliar
territory and turn ambiguity into a useful next step. We also want kind, low-ego colleagues who
are direct but constructive, generous with knowledge and good to work with. Brilliance is not a
licence for poor behaviour. People should strengthen the team, not make it harder to work in.
The roleThe role sits within Nimble’s Frontier AI Business Unit. While you will be working with a remote
client, you must be based in either our Leeds or Manchester office.
As an AI Engineer, you will be trusted to deliver complete AI features or small services independently, from integration through release and support. Your work will span the core disciplines required to build secure, useful AI products:
- Product framing
- Data and retrieval
- Application & context engineering
- Evaluation
- Reliable operation
- Build and integrate production AI services, select suitable models, and configure parameters while understanding the trade-offs between static context length, latency, and output costs.
- Build and test functional retrieval-augmented generation (RAG) pipelines that respect source permissions, handling document parsing, retrieval filtering, citation generation, and context caching.
- Implement agentic patterns—such as task routing and stateful tool execution—within production services, ensuring safe execution through deterministic fallbacks and retry limits.
- Set human approval points for higher-risk automated actions and implement programmatic guardrails to mitigate prompt manipulation and sensitive data leakage.
- Implement automated, repeatable evaluation checks for feature accuracy, response latency, and token cost to validate prompt and agent updates before deployment.
- Instrument AI applications with structured logging and tracing to capture latency, error rates, and token spend per call.
- Help define, size, and break down delivery tickets, evaluating whether a proposed feature delivers tangible user value relative to its cost, latency, and failure rate.
- You are independently delivering robust, complete AI features, ensuring non-deterministic failures degrade gracefully for the user.
- You have implemented automated evaluation checks and programmatic guardrails that the squad relies on to deploy updates safely.
- You are collaborating effectively across disciplines, consistently documenting patterns for the team, and explaining technical decisions and AI constraints clearly in reviews.
- You have established a strong working cadence with product, data, and security colleagues to test systems before release and investigate any failures in use.
- Has SC Clearance or eligibility for clearance.
- You are curious, low-ego, willing to challenge assumptions, and able to explain technical decisions clearly to engineers and clients.
- Proven ability to build robust AI features including tool calling, streaming responses, state management, and error handling.
- Hands‑on…
(If this job is in fact in your jurisdiction, then you may be using a Proxy or VPN to access this site, and to progress further, you should change your connectivity to another mobile device or PC).
View / Apply for Jobs
Matching My Jurisdiction