Principal AI Engineer — Agent Swarms & Autonomous Systems

**Principal AI Engineer — Agent Swarms \& Autonomous Systems

Manadrin and English speaking for this role is a bonus

LEC AI · London · Full-time · Office-based

Location and commitment: This role is based in our London office five days a week.

---

Read this first

This is a builder's role at the highest level of a fast-moving field. We are hiring the person who architects and personally builds our agent systems — not a manager of them, not a strategist for them. You will write production code every week, and you will be measured by what ships and what it earns. If that excites you, read on.

## About LEC AI

LEC AI is the artificial intelligence business within London Export Corporation, a London-headquartered group with seven decades of operating history between the UK and China, spanning robotics, consumer brands and media. We are not a research lab and not a startup burning someone else's money: our AI systems operate real companies, with real revenue as the measure of what works.

## The mission

We are building an organisation that runs on agents — not a single department, but the full structure: every business function, and the functions within each function, staffed by teams of specialist AI agents. Analysts, marketers, financial minds, editors, operators, researchers — organised into departments of agents with departments beneath them, working against living knowledge bases that are continuously fed from both our own companies' data and knowledge harvested from the outside world. They coordinate in structured swarms, debate and verify each other's conclusions, propose plans that humans approve, and learn permanently from the outcomes. Agents that can create, test and improve other agents under proper governance — and agents that write, run and rework their own code inside sandboxed, isolated environments with hard containment, so the system's capability grows continuously and safely rather than waiting on engineering time.

The world's frontier has just demonstrated what coordinated agent swarms can do at scale. Our ambition is the commercially disciplined version: swarms that run businesses.

You will be the person who builds this — the architecture, the coordination, the memory, the learning loops, and the governance that keeps it safe and auditable.

## What you will build and own

• Swarm-scale multi-agent architecture: many specialist agents in structured squads and hierarchies — a full organisational structure of agent departments — coordinating over durable infrastructure with defined roles, handoffs, reconciliation and containment
• Sandboxed self-coding: environments where agents safely write, execute, test and rework their own code and tools — fully isolated, permission-gated, auditable, with every change versioned and reversible
• Expert agents modelled from real-world knowledge: turning the recorded thinking of genuine specialists — documents, transcripts, recordings, decisions — into agents whose knowledge lives in files the system reads, extends and rewrites as understanding improves, continuously refreshed by harvesting what those fields' best minds are saying now
• Self-extension under governance: mechanisms by which agents create, evaluate and improve other agents and their own instructions, with every change gated, versioned and reversible
• Judgement systems: panels of agents that assess work independently, debate with evidence, and produce calibrated verdicts that become training data
• The operational spine: cost governance, quota management, human approval gates at every boundary where the system touches the outside world
• The agents themselves, hands-on: this is not a role directing a team of engineers — you personally design and build the agents, with junior support if useful. The agents are the workforce; you are their architect and builder

## What we are looking for — evidenced, not claimed

• You have designed and shipped production multi-agent systems — real users or real operations, not demos or notebooks — and can explain their architecture, handoffs and failure modes from experience
• You have worked at swarm scale or close to it: systems of many coordinating agents, or deep experience in multi-agent research and its engineering (multi-agent reinforcement learning, open-endedness, automated design of agentic systems) combined with real shipping ability
• You have built, or can demonstrably design, self-extending systems — agents that create and improve agents under governance. This is the strongest signal we can see
• You have turned real-world expertise into working agents — knowledge modelling, retrieval, memory architectures, and learning loops where corrections genuinely stick
• Reliability engineering in your bones: retries, idempotency, queues, reconciliation, containment — you know how these systems fail because yours have failed and you fixed them
• Fluency with the current open-source agent ecosystem; hands-on experience with the leading Chinese open-source models and frameworks is a strong advantage, and professional fluency.
• You have built or worked with sandboxed code-execution environments for AI systems — isolation, permissions, resource limits, rollback — and treat containment as an architectural requirement, not an afterthought
• You ship production code daily; you would rather build the system yourself than direct others to

## Who this is for

We are looking for hungry people at the top of this field — researchers who build, or builders who read the research. The one thing we offer that neither big-company labs nor academia can: your agents will run real businesses, and real revenue will be the fitness function. Full ownership, direct line to the Group CEO, our own infrastructure, and a group of companies as the laboratory.

## How to apply

Email talent@lecai.ai with your CV, links to systems you have personally built, and anything further you want to tell us — public code and papers welcome. Our process is designed to complete within two weeks and includes a practical exercise and a live walkthrough of your own work.**