Senior Software Engineer - Agentic AI
Posted 4 days 9 hours ago by Traydstream
Permanent
Full Time
I.T. & Communications Jobs
Not Specified, United Kingdom
Job Description
Overview 
Responsibilities and experience expected for a senior software engineer role focused on production-ready AI systems, high-throughput APIs, and distributed workflows.
Responsibilities- Building production-ready AI-powered systems using Python and http (FastAPI)/PubSub
- Implementing agentic AI workflows and complex agent orchestration patterns
- Rapidly shipping features while maintaining code quality and test coverage
- Deploying and managing services in Kubernetes environments
- Integrating monitoring, logging, and debugging tools for AI systems
- Optimization of agent performance, costs, and latency in production
- 6+ years of software engineering experience, with deep Python expertise including type hints, data classes, and modern Python patterns
- Production experience with FastAPI - building, deploying, and maintaining high-throughput APIs with proper error handling, middleware, and async patterns
- Hands-on experience shipping agentic AI systems using LangChain/ LangGraph (Or similar frameworks) - not just side projects, but actual production systems handling real workloads
- Strong Kubernetes experience - general understanding, debugging pods, managing resources, and using Telepresence for local development (NB- we have an infrastructure team for production management)
- Proven ability to ship code quickly in startup environments - comfortable with changing requirements, making pragmatic trade-offs, and iterating fast
- Demonstrated commitment to code quality - comprehensive test coverage, CI/CD pipelines, code reviews, and refactoring legacy code while maintaining uptime
- Real-world experience with vector databases (Pinecone, Qdrant, pgvector) including indexing strategies, similarity search optimization, and hybrid search
- Production experience managing LLM costs and latency - implementing caching layers, routing between models, handling rate limits and retries
- Built event-driven systems at scale using Kafka/RabbitMQ/Redis Streams with proper error handling and dead letter queues
- Experience debugging distributed systems - tracing requests, analyzing logs, performance profiling, and fixing production incidents
- UK/EU timezone
- Familiarity with trade finance or document processing domains
- Knowledge of prompt engineering and versioning best practices
- Experience with gradual migration of legacy systems
- Understanding of security considerations for AI systems (prompt injection, data isolation)