4 of 4 State Machine Jobs in London

documented gates, and that extends to adjacent workflows without a rebuild. Key Responsibilities Agent Orchestration Design: Own the orchestrator agent that holds the deal state machine and routes work to one specialist agent per stage, defining stage handoff protocols and cross-stage integration. Tool-Use Architecture: Design … recent hands-on experience architecting LLM or multi-agent systems. Agent Systems Expertise: Deep, hands-on experience with agent orchestration, tool and function calling, state management, and retrieval, ideally in production rather than prototypes. Technical Breadth: Expert Python (or equivalent), API and connector design, typed tool and interface design ...

model design for e-commerce concerns and understand how to handle locking sensitive resources such as a credit card charge and concepts such as state machines. Expert experience with PostgreSQL, including complex joins and reporting queries, and familiarity with data modeling and performance optimization. Familiarity with cloud infrastructure (e.g. ...

liquidations, and collateral releases Plug into custodians, market data providers, and other financial infrastructure Build guardrails that make unsafe states genuinely hard to reach (state machines, validation, audit trails, reconciliation) Write tests that catch the scary stuff: integration, property based, simulations, fuzzing Own systems end to end, from first … mission critical Deeply gets concurrency, race conditions, distributed systems, idempotency, ordering, and partial failure Has a strong testing and correctness mindset Can design safe state transitions, audit trails, and reconciliation flows Owns technical design even when the product is messy and undefined Has Rust experience, or strong evidence they ...

think about until it breaks: making sure the thing underneath never does. What You'll Accomplish Own the orchestration engine: scheduling, execution, state, retries, concurrency and cancellation, the machinery that decides what runs, when, and what happens when it doesn't. Python and FastAPI on AWS, shipped to production … real architecture calls, and know why certain patterns exist rather than just following them Judgment about distributed systems: retries, idempotency, at-least-once delivery, state machines, queues, race conditions that only show up under load. You've debugged the sort of failure that only happens in production ...