Senior Machine Learning Engineer (AI / LLM Systems)
Remote (United Kingdom) | Permanent | Flexible compensation + equity
We’re working with a well-established, tech-led business that is building a new AI product focused on real-world tasks, workflows, and decision-making.
This is a small, high-calibre team building systems where model capability is transformed into reliable, production-grade ML systems, with a strong emphasis on ownership, iteration, and real-world performance.
The product focuses on:
-
Long-running AI workflows
-
Persistent context across interactions
-
Multi-step reasoning and task execution
-
Integration with external tools and systems
The core challenge is designing ML systems that can behave reliably in production, even when model outputs are inherently non-deterministic.
The Role
This role sits at the core of the ML layer powering the product.
The focus is on designing and operating systems that enable:
-
ML models to run reliably in production
-
End-to-end pipelines from training to inference
-
Continuous evaluation and iterative improvement
-
Systems that perform consistently under real usage conditions
You’ll be working on:
-
Training, inference, and evaluation pipelines
-
LLM-based systems and agent-style workflows
-
Debugging model behaviour using real-world signals
-
Optimising performance across latency, cost, and reliability
-
Production monitoring, logging, and system stability
What They Care About
The hiring bar is centred around real production ML experience:
Whether you have:
-
Shipped ML systems used by real users
-
Owned ML systems end-to-end in production
-
Worked with modern LLMs beyond simple API integration
Your exposure to practical challenges such as:
-
Model behaviour debugging and failure analysis
-
Latency, throughput, and cost optimisation
-
Monitoring, observability, and evaluation frameworks
-
Scaling ML systems in production environments
Tech Environment
-
Python (core ML and backend language)
-
PyTorch / modern ML frameworks
-
LLM ecosystem (OpenAI, Anthropic, etc.)
-
GPU-based training and inference
-
Docker and Kubernetes
-
AWS, Azure or GCP
The emphasis is on how you build and operate ML systems, rather than specific tools.
Team & Working Style
-
Fully remote-first, work from anywhere
-
Small, highly capable engineering team
-
Strong emphasis on ownership and delivery
-
Fast iteration cycles with real user feedback
-
Comfortable working in evolving systems and making pragmatic decisions