AI Engineer -- Agentic AI Systems
19 hours ago
London
AI Engineer -- Agentic AI Systems LEC AI Central London (Knightsbridge) | Full-Time | 5 Days On-Site About LEC AI LEC AI is the artificial intelligence division of London Export Corporation, a British trading group established over 73 years ago operating across multiple sectors and international markets. We are building production AI systems that serve as the operational brain for the organisation -- multi-agent architectures that handle memory, reasoning, task management, document intelligence and cross-departmental coordination in real time. This is not a research lab. We ship production systems, iterate fast, and operate with the intensity of a startup backed by the resources and commercial reach of an established international group. The Role We are looking for an AI Engineer with 2-5 years of engineering experience who can build end-to-end: from designing agentic architectures to deploying them on cloud infrastructure, from writing RAG pipelines to wiring up tool-use protocols, from prompt engineering to system design. You will work directly with senior leadership and the founding technical team. This is a small, high-trust environment where your work ships to production the same week and is used daily by the people running the business. This role is for someone who thrives in ambiguity, moves fast when requirements change daily, and would rather build something that works today than plan something perfect for next quarter. You will be a core builder -- not a ticket-taker. You will own entire systems, make architectural decisions, and ship directly to production. If you need someone to tell you what to do every morning, this is not the role for you. Key Responsibilities Agentic AI Systems • Design, build and maintain multi-agent architectures with tool calling, memory, reasoning loops and inter-agent communication, • Implement agentic loops (prompt assembly, LLM call, tool execution, iteration) with circuit breakers and observability, • Build and extend tool registries, function calling schemas, and execution pipelines, • Work with tool-use protocols like MCP for interoperability across agents and services RAG & Memory Systems • Build and optimise retrieval-augmented generation pipelines across structured and unstructured data, • Implement hybrid memory stacks using things like vector databases, knowledge graphs, session caching and object storage, • Design embedding pipelines, chunking strategies, and semantic search with re-ranking, • Build document ingestion systems that process PDFs, spreadsheets and unstructured documents into searchable knowledge Real-Time & Multi-Modal Systems • Build streaming data pipelines, live communication channels and event-driven architectures, • Work with audio processing, speech-to-text and multi-modal data pipelines, • Design systems that handle both batch and real-time workloads in production Software Engineering & System Design • Write production backend and frontend code (like Python, JavaScript/TypeScript, React), • Design APIs, data models, and system architectures that scale, • Understand trade-offs between simplicity and abstraction -- build what is needed, not what is theoretically elegant, • Write code that other people can read, debug and extend, • Implement production security practices -- access control, data isolation, PII handling and authentication flows Deployment & Infrastructure • Deploy and manage containerised services on cloud infrastructure, • Set up CI/CD, monitoring, logging and observability, • Manage databases and storage in production (like Postgres, graph databases, caches, S3-compatible stores), • Handle DNS, reverse proxies, SSL, tunnels -- the full stack of getting things live and keeping them live Rapid Prototyping & AI-Assisted Development • Use AI-assisted development tools effectively to move at speed, • Prototype new capabilities quickly, validate with real usage, then harden, • Comfortable building MVPs in hours, not weeks Candidate Profile The ideal candidate has 2-5 years of engineering experience and is a builder who ships. You combine deep AI/LLM knowledge with real software engineering discipline and the operational grit to deploy and maintain production systems. Must have: • 2-5 years of software engineering experience, with meaningful time spent on AI/LLM systems, • Hands-on experience building agentic AI systems (not just calling an API and returning the response), • Strong understanding of RAG architectures, vector databases and semantic search, • Production software engineering skills -- you have deployed and maintained real systems, • Experience with containerisation, Linux, and cloud deployment, • Comfort with rapidly changing requirements and ambiguity, • A bias toward action -- you try things, break things, fix things, ship things Strong signals: • You have built multi-agent systems or AI orchestration pipelines, • You understand tool-calling protocols, MCP, and LLM integration patterns, • You have designed memory or knowledge management systems for AI, • You have experience with knowledge graphs, • You have built agentic evaluation frameworks -- measuring agent accuracy, tool-use correctness, hallucination rates, task completion, and regression testing across prompt or model changes, • You contribute to or maintain open-source AI tooling, • You have built full-stack applications with AI backends, • You can explain system design trade-offs clearly and make pragmatic choices What we are NOT looking for: • Pure researchers with no production deployment experience, • Engineers who need stable, well-defined requirements to function, • People who over-architect and under-deliver, • Anyone who thinks AI engineering is just prompt engineering What Success Looks Like Within the first 6 months, you will have: • Built and shipped a complete new agent or system end-to-end -- something with a measurable output that the business uses, • Taken ownership of core AI systems and delivered tangible improvements to production, • Picked your own priorities from the roadmap and executed with full autonomy, • Deployed new integrations or capabilities that are live and running without hand-holding, • Contributed original ideas to the architecture, not just implemented what was handed to you Why Join This is a ground-floor opportunity to build the AI infrastructure of an international trading group. You will not be maintaining someone else's system -- you will be architecting and building it. This is a founding-stage AI team inside an established international group. You will not be engineer #200 maintaining someone else's decisions. You will be one of the first builders shaping the architecture, the culture, and the direction of the AI division from the ground up. The pace is startup. The backing is institutional. The problems are real and commercially meaningful. As we grow, early team members will grow with us -- into technical leads, heads of engineering, or whatever shape makes sense based on what you build and how you build it. If you want to build AI systems that actually run a business -- not demos, not prototypes, not pitch decks -- and you want the kind of ownership and trajectory that only comes from being there early, this is the role. How to Apply Send us your CV along with links to practical work you have done -- GitHub repos, deployed projects, open-source contributions, technical blog posts, or anything that shows what you have built. We care far more about what you have shipped than where you studied or what certifications you hold. A strong portfolio of real work is the single best indicator we use to shortlist candidates. If you have built something interesting, show us.