Senior Applied AI Engineer (Private Sector Contractor)

Role: Senior Applied AI Engineer (Private Sector Contractor)

Rate: £700–£900 per day, dependent on experience

Location: Hybrid: 2 days a week on site at Harwell, Oxfordshire; remainder remote (UK-based)

Engagement: Contract, outside IR35: initial (6)-month term with strong likelihood of extension

Start: Immediate Start

A note from the Founders

Oxford Dynamics is at an inflection point.

We build frontier AI for environments where the cost of a wrong decision is high: defence and national security, critical infrastructure, and commercial organisations making high-stakes decisions on complex data. We are growing fast, and the constraint on how fast is no longer ideas or opportunity. It is whether the systems we ship are good enough for the people who stake decisions on them.

That is where you come in. We need engineers who can build the systems this mission depends on, and build them to a standard that holds when the answer has to be right. You will be trusted to own real systems end-to-end, held to a single high standard, and judged on one thing above all: whether what you ship earns the confidence people place in it when the decision matters.

If you are the kind of engineer who builds rather than assembles, who wants your work measured by whether it holds up in production rather than whether it demos well, and who wants to see the direct line from your code to systems people depend on, we would love to hear from you.

Who We Are

Founded in 2019, Oxford Dynamics is a fast-growing UK frontier technology company developing both digital and physical AI systems built to operate in dynamic, mission-critical environments.

At the core of everything we build is AVIS™ (A Very Intelligent System), our orchestration platform: the intelligence layer that fuses multi-modal data, including text, imagery, telemetry and sensor feeds, so operators can interrogate complex information at speed and make better decisions under pressure. AVIS™ powers our digital decision products, such as ORION for multi-source decision-to-effect, and it powers our physical platforms.

We work where wrong decisions can be catastrophic. We partner with defence, security and private sector organisations internationally to bring frontier AI into the environments where it matters most, whether that is protecting nations, infrastructure and lives, or transforming how commercial organisations reason over their data and make high-stakes decisions.

As a Senior Applied AI Engineer at Oxford Dynamics, you will build and ship agentic AI systems in production: designing the multi-agent pipelines, the reasoning and retrieval, the model integration, and the infrastructure that carries them into production. These are systems people rely on to make consequential decisions, so what you build has to earn its confidence honestly and hold up under scrutiny.

This is a builder's role, not an integrator's, and it is unambiguously senior. We need an engineer who designs and implements systems rather than wiring together off-the-shelf parts, who takes technical ownership of the workstream and sets the direction other engineers build to. You will make systems-level design decisions, write production-grade code, and own the hard engineering that makes agentic AI reliable in demanding production environments, including secure, on-premises and offline deployments, from architecture through to deployment.

That ownership is real. You will own features end-to-end and be accountable for how they behave under pressure, not just whether they work on your machine. Reliability, latency and failure behaviour are your concern rather than an afterthought, because in the environments we deploy into, reliability and trust matter more than speed to market.

Frontier here means dependable, not experimental. You need the engineering depth to build systems that keep working under real-world conditions, whether that is heavy load, degraded connectivity or a fully offline environment, and the judgement to know the difference between a demo and a system someone can stake a decision on.

Key Responsibilities

• Build the systems themselves. You design and build multi-agent AI systems from the ground up, using frameworks such as LangGraph or Haystack where they help, and building the orchestration, state and tool-calling infrastructure underneath them when they do not.
• Engineer for production. You write clean, tested, maintainable code and make sound systems-level decisions, treating reliability, latency and failure behaviour as first-class concerns.
• Put models to work. You integrate large language and vision-language models into real reasoning, search, summarisation and task-execution pipelines, and build the retrieval, memory and guardrails that make them dependable.
• Deploy where it is hard. You take systems into cloud, on-premises and, where required, fully offline environments, and engineer for the security and reliability constraints that come with each.
• Own the full pipeline. You build production pipelines from data ingestion through to inference, owning the whole path rather than a slice of it.
• Make confidence earned, not asserted. You build the evaluation, observability and guardrails that show how a system actually behaves, its agent behaviour, model performance and failure modes.
• Set the technical bar. You act as a technical authority for agentic AI, taking technical ownership of the workstream, setting the design patterns and direction other engineers build to, and leaving behind systems and documentation the team can maintain after your engagement ends.

What We're Looking For

• Significant professional software engineering experience (typically 6+ years), including several years building LLM-based or agentic systems, with a demonstrable track record as a senior or lead engineer.
• Strong software engineering fundamentals: you write clean, tested, production-grade code and make sound systems-level design decisions, not just working prototypes.
• Commercial experience building multi-agent or agentic AI systems in production, owning them as an engineer rather than integrating them as a user.
• Strong Python, with hands‑on experience building on LLM frameworks (LangGraph, LangChain, Haystack, or similar) and the ability to work below the framework when you need to.
• Experience designing and deploying AI/ML systems into real production environments, not just notebooks or demos.
• A solid grasp of the engineering around models: inference, latency and performance, data pipelines, state and memory, and tool or function calling.
• Hands‑on experience building search and retrieval over very large multi‑modal datasets (3TB+ of text, imagery, telemetry and sensor data): indexing, embedding and querying at that scale, and feeding the results into LLM or agentic pipelines with the latency and precision they need to be useful.
• Fluency with Docker, Git and cloud platforms (AWS preferred).
• Understanding of secure deployment patterns: air‑gapped, on‑premises or sovereign cloud.
• A builder's mindset: you ship real, dependable systems and can explain them clearly to technical and non‑technical people alike.
• Senior technical ownership: you own your delivery end‑to‑end, take technical ownership of the workstream, and set the direction other engineers build to. You are disciplined about documentation and knowledge transfer so what you build outlives your engagement.

Nice to Have

• Experience with multimodal reasoning
• Experience with edge or offline AI deployments
• Familiarity with Kubernetes (EKS/OpenShift) for managing deployed applications
• MLOps experience: model evaluation, monitoring, reproducibility
• Observability tooling for agentic systems (model drift, agent behaviour, performance monitoring)
• Experience with agent orchestration patterns and inter‑agent communication protocols (e.g. A2A)
• Familiarity with the Model Context Protocol (MCP) for tool and context integration in agentic systems
• Familiarity with secure‑by‑design development principles (ISO 27001, NIST, OWASP)
• Experience in defence, national security, or similarly regulated environments
• Contributions to open‑source AI/ML projects

Why This Role?

This is a role for engineers who want to build things that matter and see where they land. You will own real systems end-to-end, work at the frontier of agentic and generative AI, and see your work deployed with organisations where getting the answer right matters. Few engineering roles let you say that the thing you built is trusted by people making consequential decisions. If you build to a high bar and want your work to count, we would love to meet you.

Why Oxford Dynamics?

Join one of the most exciting, fastest‑growing areas in the world: the convergence of AI and robotics. Every member of the Oxford Dynamics team has a major impact on the products and services we provide. Regardless of job title, you will make a real difference and learn from colleagues across every area of our business.

What We Offer

• Competitive day rate, dependent on experience
• Genuine technical ownership and the scope to shape architecture, not just deliver tickets
• The opportunity to work with a fast‑growing, successful early‑stage business, with a realistic prospect of extension or a longer‑term arrangement
• Flexible and hybrid working pattern

Oxford Dynamics is committed to creating an inclusive team experience for all. Regardless of race, gender, religion, sexual orientation, age, disability, or parental status, we believe our work is at its best when everyone feels free to be their authentic self.

#J-18808-Ljbffr