Senior Machine Learning Engineer

Accurx is where conversations happen with and about patients. For decades, the NHS has struggled with fragmented systems that make simple tasks feel impossible. We’re changing that by building a single, system-wide platform that connects everyone through communication. What started as a way for GPs to text a patient has now evolved into an all-in-one digital toolkit used by 98% of GP practices. Our platform now powers Total Triage to manage patient demand, and Self-Book, which lets patients schedule their own appointments in seconds. We’ve automated routine care with Patient Questionnaires for long-term conditions, while Accumail finally allows staff-to-staff communication to happen instantly across different care settings. We’re now pushing the boundaries of the consultation itself with Accurx Scribe, our AI-powered note-taker that drafts medical notes in real-time.

The team is a mission-driven group of 80 engineers based in London and the surrounding areas, united by the challenge of fixing healthcare communication. We are innovating for the NHS at a scale and depth that has never been done before, solving the real-world problems that stand between millions of patients and the care they need. We are a highly cross-functional group where engineering, product, data, and security collaborate as true peers. We foster a low-ego, high-impact environment that values expertise and new ideas, maintaining the high standards required to build and scale a national healthcare communication platform.

As a Senior Machine Learning Engineer, you'll join our Triage Intelligence team – a high-impact, cross-functional group of software engineers, Data Scientists, clinicians, and a Principal Product Manager. The team builds the intelligence that powers how Accurx understands healthcare communication across the entire product suite, with a current mission to interpret patient healthcare requests and enable fast, informed decisions about the next step in a patient’s care.

Challenges you’ll solve:

• You will take ownership of the end-to-end technical design of ML systems - from data ingestion and training pipelines through to deployment and real-time monitoring - ensuring AI features are performant, observable, and easy to debug.
• You will act as the primary technical partner to our Data Scientists, evolving experimental code into high-quality, extensible library modules and building reusable components that improve developer efficiency and reduce cycle times across the team.
• You will translate complex clinical and product objectives into efficient computational tasks, selecting the right tools - whether Deep Learning, NLP, or off-the-shelf - and making smart build-vs-buy trade-offs that align with Accurx's long‑term technical goals.
• You will help turn one-off experiments into repeatable platform capabilities, building models and systems that can be adopted by other Accurx product teams and proactively resolving cross-team dependencies as new, more efficient AI models are rolled out.

You should apply if:

• You have extensive experience with a variety of ML techniques (e.g. Transformer-based NLP, Deep Learning, Tree-based methods, Bayesian modelling) and the ability to select the right tool, not just the trendiest one, for a clinical problem.
• You have a proven track record of taking models from a Jupyter notebook to a high-availability production environment, managing everything from data versioning to model serving.
• You have mastery of a production-grade language (e.g. Python, C#, or Go), with a focus on writing extensible, modular libraries that other teams can adopt.
• You have experience defining both offline metrics (how the model performs in training) and online metrics (how the model impacts real-world user behaviour and triage outcomes).
• You bring collaborative rigour to design and code reviews, surfacing erroneous assumptions before they reach production, and can preemptively identify technical risks — such as data leakage or hardware constraints — before they jeopardise the product roadmap.

Bonus points if you have:

• Experience working with sensitive, regulated data (GDPR, HIPAA) and implementing privacy-first techniques like differential privacy or secure data handling.
• Familiarity with infrastructure-as-code tools like Terraform, Kubernetes, or MLOps platforms such as SageMaker or Vertex AI.
• Hands‑on experience fine‑tuning LLMs, prompt engineering, or managing the latency and cost trade-offs of large language models at scale.

What’s in it for me:

• You’ll be joining an established but fast-growing Tech for Good movement, where we’re led by our Principles and our mission to fix healthcare communication.
• Benefits to suit you: adjust your healthcare cover, your pension or life insurance, whatever stage you’re at in life.
• Flexible working: We are an office first culture and ask that you’re in our (dog-friendly) Shoreditch office 3 days a week, with core hours of 10am - 4pm.
• Time off: You’ll get 28 days of holiday (plus bank holidays) and up to 4 weeks to work from anywhere per year.
• Family matters: We offer enhanced parental leave, fertility support and parental loss support.
• We have our very own Chef! Free healthy breakfasts, snacks and lunches will be provided, with the occasional sweet treat!