Head of AI

About Dyad

Dyad is focused on enhancing healthcare delivery and operational efficiency through innovative data management platforms. Their technology combines semantic AI, knowledge graphs, and generative AI to improve workflows for healthcare providers and payers primarily in the UK and US. Operating as a vibrant early-stage startup with about twenty employees, Dyad encourages impactful contributions that shape both their business trajectory and organizational culture.

Product Overview

The company's core offering, the Dyad Platform, integrates advanced AI technologies to create explainable intelligence tailored for healthcare operations. Their product suite includes BetterLetter, an AI solution designed to assist GP practices by easing administrative workloads related to clinical correspondence, thereby enhancing clinical coding, task follow-up, workflow efficiency, and audit performance.

Role Summary

Dyad is looking for a Head of AI to spearhead the design and execution of their graph-integrated generative AI systems. This senior, hands-on leadership role focuses on guiding a team responsible for developing production systems that process unstructured clinical text, structured knowledge bases, and generative AI methods. The candidate is expected to convey complex AI concepts clearly to various stakeholders and propagate a culture of continuous learning and knowledge sharing within the team.

Key Responsibilities

• Lead the architecture, design, and end-to-end ownership of AI systems integrating NLP pipelines, LLM-based reasoning, knowledge graph grounding, benchmarking, and validation frameworks.
• Establish and maintain standards for system design across the Applied AI stack ensuring robustness, production readiness, and alignment with product goals.
• Drive daily technical coordination by translating product requirements into coherent plans, aligning multidisciplinary AI components, and sustaining architectural consistency as the product evolves.
• Develop and uphold evaluation frameworks addressing hallucination detection, clinical concept extraction metrics, regression testing, schema-constrained generation, and iterative human-in-the-loop feedback to improve system accuracy and reliability.
• Embed compliance by designing AI workflows with traceability, auditability, and data minimization, aligning with medical device regulations and data protection laws in the UK and US, collaborating with Clinical Safety and Quality Assurance teams to mitigate architectural risks.

Qualifications and Expertise

• Master's degree in Computer Science with a focus on AI or equivalent experience with 5+ years delivering production AI/NLP systems, including technical leadership or architecture roles.
• Proven ability to design production AI systems integrating large language models with structured knowledge, balancing symbolic reasoning, probabilistic inference, and pattern matching techniques.
• Experience developing NLP pipelines combined with structured data validation or knowledge graph applications, with specialized expertise in clinical NLP, entity recognition, and coding standards like SNOMED CT, ICD, UMLS.
• Familiarity with document AI methods including OCR, layout-aware or multimodal architectures.
• Strong proficiency in Python for NLP, LLM orchestration, evaluation tooling, and data processing; experience with Elixir in production is a plus or openness to learning.
• Hands-on in prompt engineering, schema-constrained generation, evaluation and benchmarking of LLMs, model versioning, regression testing, and iterative improvements.
• Skill in integrating AI pipelines constrained or validated by knowledge graph structures, collaborating effectively with knowledge engineers, and understanding scaling challenges.
• Preferred experience in regulated or high-assurance environments, balanced with an ability to operate in fast-paced startup settings demonstrating ownership and accountability.

Personal Traits

• Systems thinker prioritizing coherence over novelty.
• Focus on pragmatic, production-ready AI solutions rather than purely experimental research.
• Strong ownership mentality with clear communication and documentation skills to disseminate architectural knowledge.

Benefits

• Competitive salary package.
• Company pension scheme.
• 25 days of paid vacation annually, prorated.
• Flexible and enjoyable hybrid working environment.
• Access to an Employee Assistance Programme for health and wellbeing.
• Modern, dog-friendly office near Chancery Lane with complimentary beverages.