SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Limbic is building an AI therapist that delivers end-to-end Cognitive Behavioural Therapy (CBT) to patients in the US and UK under clinical supervision. The company has supported over 820,000 patients and is deployed across 50% of NHS Talking Therapies services. Limbic holds UKCA Class IIa medical device certification and is the first AI-led mental healthcare company selected for the FDA's TEMPO pilot. Recent research published in Nature Medicine showed their AI agents scored higher than licensed CBT therapists on standard therapy skill measures.
As an AI/ML Engineer, you will work at the intersection of computational psychiatry, clinical science, and software engineering. You'll join the product team building a voice agent that delivers structured, evidence-based CBT over the phone to adults with depression and anxiety, with licensed clinicians overseeing every patient's care. The product is live and growing, so your work reaches real patients quickly.
Key responsibilities include:
- Improve how the AI therapist runs sessions, from session structure and case formulation to therapeutic techniques and continuity across treatment courses, working closely with clinicians
- Strengthen real-time safety systems, including risk detection and agent responses when risk is disclosed
- Build evaluations and LLM-as-judge autolabellers validated against clinician labels to quickly assess whether changes improve care
- Make voice conversations feel natural: responsive, well-paced, and able to recover gracefully when conversations go off script
- Help patients stay with treatment through AI features such as session reminders and clear therapy expectations, tested via experiments
- Optimize collection of outcome measures (PHQ-9, GAD-7) including methods, timing, and predictive modelling of recovery
- Ship changes safely through incremental rollouts, automated testing, and clinical sign-off processes built for regulated medical devices
You will own problems end-to-end: frame them with clinicians and product managers, ship changes, measure impact, and iterate. Your work is built on shared platforms for agents, evaluation, and safety, so solutions are reused across Limbic.
Requirements:
- Educational background: Either (1) PhD in Computational Psychiatry or related field with demonstrated industry experience in Data Science, Product, or ML engineering; OR (2) BSc/MSc in Computational Psychiatry or related field (research focus) plus 2+ years of ML engineering or Data Science experience in an Engineering/Product Development environment
- Proficiency in Python or TypeScript/JavaScript
- Experience building LLM-powered features and track record of eval-driven development
- Hypothesis-driven work approach with ability to balance product development and scientific rigour
- Production ML experience or high-impact scientific publications (predictive modelling, reinforcement learning)
- Experience modelling text/language data, including LLM classifiers and vector-based methods
- Experience modelling longitudinal data and working with databases (SQL or NoSQL)
- Strong understanding of evidence-based mental health care (CBT) and deep interest in causal mechanisms of therapeutic benefit
- Bias to action: break big problems into pieces shippable within a week; comfortable with incomplete information
- Proficiency with version control systems, preferably Git
- Ideal but not required: Experience building LLM agent harness or real-time voice AI
- Ability to work from London (Spitalfields) office on Tuesdays and Thursdays, with flexibility for remote/office work other days
