12 days ago
London, United KingdomSenior
Responsibilities
- Design and build LLM-powered systems using RAG, fine-tuning, tool use, and multi-agent orchestration.
- Develop agentic workflows for automation, reasoning, and conversational experiences.
- Define autonomous and human-in-the-loop system patterns.
- Build and scale solutions on the Azure AI stack, including Azure OpenAI, Azure AI Studio, Azure Machine Learning, and Cognitive Services.
- Create reusable prompt orchestration layers, vector search and retrieval pipelines, and evaluation and observability frameworks.
- Design and implement offline and online LLM evaluation frameworks.
- Implement AI safety guardrails covering hallucination control, filtering, explainability, and content safety.
- Partner with Trust & Security to embed AI risk controls by design.
- Build reusable AI capabilities such as stylist reasoning, product understanding, and copilots.
- Enable horizontal reuse across squads by building capabilities once and scaling them broadly.
- Optimize systems for performance, cost, token usage, and reliability.
- Champion AI engineering, governance, platform thinking, and inclusive responsible AI development.
Requirements
- Strong hands-on experience with the Azure AI ecosystem, including Azure OpenAI, Azure AI Studio, and Azure Machine Learning.
- Experience building and operating production-grade AI systems on cloud infrastructure, including APIs, event-driven architectures, and scalable compute such as AKS or similar.
- Proficiency in Python and experience with modern AI frameworks such as Semantic Kernel, LangChain, or equivalents.
- Deep understanding of LLMs and transformer-based models, including embeddings, tokenization, and context management.
- Experience designing and optimizing LLM-powered systems, including fine-tuning approaches, structured prompting, and model selection trade-offs.
- Experience building end-to-end RAG pipelines involving retrieval strategies, vector search, and grounding techniques.
- Experience designing agent-based systems with multi-agent patterns, tool usage, and workflow orchestration.
- Understanding of state, memory, and event-driven pipelines in conversational or decisioning systems.
- Experience implementing evaluation frameworks to measure model quality and performance.
- Strong understanding of AI safety, governance, and guardrails, including hallucination mitigation, content safety, and explainability.
- Experience designing secure, scalable, and cost-efficient production systems.
- LLMOps or MLOps experience, experiment tracking, observability, and performance or cost optimization are desirable.
Benefits
- Employee discount and access to employee sample sales.
- 25 days of paid annual leave plus an additional celebration day.
- Discretionary bonus scheme.
- Private medical care scheme.
- Flexible benefits allowance that can be taken as extra cash or used toward other benefits.
- Personalized learning and in-the-moment development opportunities.
- ASOS is Disability Confident Committed and provides reasonable adjustments during recruitment.
Categories
