Senior AI Engineer
- Chunking and ingestion architecture with deterministic, validation-gated pipelines handling diverse enterprise document types with indexing and contextual enrichment
- Retrieval pipeline with vector search, BM25 hybrid retrieval, RRF re-ranking, embedding model selection and optimisation
- Agent orchestration designing stateful, multi-step agent workflows with LangGraph and building chains and tool-use patterns with LangChain
- Observability including tracing, debugging, and monitoring pipeline performance end-to-end with LangSmith and catching retrieval and reasoning failures before they reach users
- Model evaluation and tuning using RAGAS-based evaluation frameworks, benchmarking retrieval recall and answer faithfulness, iterating toward production quality
- Open-weight LLM deployment evaluating and shipping models (Llama 4, Qwen3, MiMo-V2) under memory and latency budgets via Amazon Bedrock and self-hosted inference
- Workflow orchestration with async pipeline design using Temporal.io and containerised deployment on Kubernetes via GitLab CI
- Strong software engineering skills in .NET and/or Python with experience building production-grade APIs and services
- Applied algorithmic problem solving including search, ranking, clustering, and data extraction/transformation pipelines
- Solid understanding of LLM architectures and trade-offs (e.g. GPT-family, Claude, Llama, and similar models)
- Experience integrating LLM and agent APIs (e.g. Azure OpenAI, OpenAI API, Claude API, Amazon Bedrock)
- Design and implementation of agent-based systems for automation, orchestration, and decision support
- Experience working with relational, document, and graph databases including data modeling and querying
- Web data ingestion experience including web scraping, schema extraction, and handling changing source structures
- Experience building reliable, scalable, maintainable AI systems including logging, error handling, and performance considerations
- Fine-tuning and preference optimization techniques (e.g. supervised fine-tuning, RL-style approaches via APIs)
- Designing pipelines that convert unstructured data into reliable structured representations
- Exposure to event-driven or background processing architectures for long-running AI workflows
- Experience contributing to or operating AI-powered systems in production