Responsibilities
- Design and implement agents
- Build agents that can plan tasks, call tools/APIs, decompose problems, and iterate based on feedback and intermediate results.
- Orchestration & routing
- Develop orchestration services that coordinate multiple models, agents, and tools (e.g., planners, executors, critics, evaluators), including policy logic and routing decisions.
- Tooling & integration
- Integrate agents with internal and external tools: APIs, databases, retrieval systems, transactional systems, and product features. Handle auth, error cases, and safety constraints.
- Context & state handling for agents
- Work closely with context management to maintain agent state across steps and sessions, including working memory, scratchpads, and long-running workflows.
- Reliability, safety & guardrails
- Implement checks, constraints, and fallback policies so agents behave predictably and safely, especially when they can take real-world actions.
- Observability & debugging
- Build tracing, logging, dashboards, and replay tools to understand agent behavior, diagnose failures, and iterate on policies and prompts.
- Experimentation & iteration
- Design experiments to compare agent architectures, policies, and tool sets; iterate based on offline and online metrics and qualitative user feedback.
Qualifications
5+ years in software engineering, ML engineering, or applied AI roles, including experience shipping systems that rely on LLMs or other ML components.
Strong programming skills in Python and experience building backend services (REST/gRPC, microservices, or similar).
Solid understanding of system design basics: latency, throughput, reliability, and failure modes.
Hands-on experience working with LLM-based applications (e.g., chatbots, copilots, automation tools, or internal agent-like systems).
Familiarity with at least one agentic or orchestration framework (e.g., LangChain, Autogen, CrewAI, custom in-house frameworks), or evidence of having built similar functionality yourself.
Ability to reason about complex multi-step flows, failure modes, and edge cases.
Comfortable diagnosing issues that span prompts, tools, infra, and product logic.
Bonus Points
Experience designing multi-agent or multi-stage workflows (e.g., planner/solver/critic architectures).
Background integrating LLMs with transactional systems, internal tools, or external APIs.
Exposure to safety/guardrail systems: policy engines, content filters, or constrained tool execution.
Familiarity with cloud platforms, containers (Docker), and observability tools (metrics, tracing, logging).
What we can offer you
Holiday purchase
Private medical
Income protection
Hybrid working model (3:2)
Attractive pension scheme
Positive work life balance
Learning and development
Life insurance
Lenovo and Motorola products discounts
Lifestyle discounts
Cycle to work
MyGymDiscounts
Mortgage advice and support
Referral bonus
Free onsite parking
We are an Equal Opportunity Employer and do not discriminate against any employee or applicant for employment because of race, color, sex, age, religion, sexual orientation, gender identity, national origin, status as a veteran, and basis of disability or any federal, state, or local protected class.
Additional Locations
United Kingdom
United Kingdom
AI PROCESSING NOTICE
We use AI-based tools to support some of our processes (e.g. online interviews recordings and transcripts) in order to achieve better efficiency, accuracy and for our documentation purposes. AI can make mistakes, but we always make sure that the outputs are manually reviewed by a human. You can always opt-out or contact us in case of any question.