Machine Learning Engineer - LLMs & Agentic AI (Lead I - Data Science)

Role description

Machine Learning Engineer - LLMs & Agentic AI

London (Hybrid - 3 days onsite). Full-time, Permanent. Join by December 2025.

We are seeking a skilled and forward-thinking Machine Learning Engineer to join our Applied AI and R&D team in London. You'll play a key role in building the next generation of agentic AI systems that enhance Oversight's enterprise spend management and risk control solutions.

This hands-on role combines deep machine learning expertise with cutting-edge work on Large Language Models (LLMs), Small Language Models (SLMs), retrieval-augmented generation (RAG), and multi-agent frameworks. You'll collaborate with data engineers, AI researchers, and product teams to design and deploy scalable, explainable, and autonomous AI solutions.

• 
  
• Design, fine-tune, and deploy ML/LLM models for production.
  
• Build and optimize RAG pipelines using vector databases and embeddings.
  
• Develop multi-agent workflows leveraging LangChain, LangGraph, or MCP.
  
• Implement prompt engineering, safety, and explainability modules in agentic systems.
  
• Collaborate with data teams on real-time pipelines, model monitoring, and drift detection.
  
• Experiment with AI/GenAI frameworks and translate research prototypes into production-ready solutions.
  
• Contribute to best practices in ML engineering, mentoring, and technical documentation.
  

• 
  
• Bachelor's or Master's in Computer Science, Machine Learning, or related field.
  
• Minimum of 3+ years experience building and deploying ML systems.
  
• Proficiency in Python, PyTorch, TensorFlow, Hugging Face Transformers.
  
• Experience with LLMs/SLMs (e.g., GPT-4, Claude, Gemini, LLaMA).
  
• Strong skills in RAG, vector databases, and data engineering (SQL, Spark, Ray).
  
• Familiarity with cloud ML platforms (AWS, Azure, or Databricks ML).
  
• Understanding of MLOps, CI/CD, and infrastructure-as-code.
  

• 
  
• Experience with agentic frameworks
  
• AI safety/guardrails
  
• Anomaly detection / open-source AI contributions