Professional Experience
Building AI ThatResearchers Trust
NHS doctor turned product leader. I saw problems on the wards that technology could address at a scale that clinical work alone couldn't, and I've spent the last decade building products to do exactly that.
I lead AI product strategy for evidence-grounded research workflows, combining retrieval, reasoning, trust mechanisms, and product design to help experts move from information overload to confident decisions. My work spans Scopus AI, LeapSpace, RAG Fusion, and multiple patents in retrieval-augmented generation.
Core Competencies
Product Skills
- Product Strategy
- TARS Framework
- Working Backwards Methodology
- Strategic Innovation Execution
- OKRs/KPIs Design
Technical Expertise
- ElasticSearch (Lexical + Semantic)
- Python
- Graph Database Design (Neo4j)
- Vector Space Modeling
AI-Powered Technologies
- RAG Architectures (3 Patents)
- Multi-Agent Systems
- LLM Evaluation & Responsible AI
- Embedding Pipelines & Clustering
Healthcare & Clinical
- Clinical Medical Practice
- Healthcare Regulation
- Patient Safety
- Medical Research
Experience
Senior Director, Product Management
In 2020, I started experimenting with GPT-2 to see whether large language models could genuinely help researchers find what they need. I built the original Scopus AI prototype, along with the RAG Fusion algorithm, focusing on creating systems that work with how researchers actually think and collaborate, not just what an algorithm can retrieve. Working with brilliant minds like Henry Cleland and Erik Schwartz, this prototype became the foundation that enabled a dedicated team to create Scopus AI.
Now I lead LeapSpace, the research-grade AI workspace built on 100M+ Scopus records and 15M+ peer-reviewed full-text articles. I authored the vision manifesto, designed the corporate pricing model, and took the product from concept through commercial launch. I also built much of the early agentic architecture hands-on: the iterative agent system, orchestrator routing, and institutional analytics capabilities that gave the engineering team a working foundation to build from.
- Built the original Scopus AI prototype that became the foundation for a platform now used by millions of researchers worldwide.
- Took LeapSpace from first commit to commercial launch, growing to thousands of monthly active users across 50+ countries within weeks of going live.
- Led the shift from keyword matching to semantic understanding across five research products, helping over 100M researchers find papers and connections they would have missed before.
- Developed RAG Fusion, a patent-pending approach to retrieval-augmented generation, with over 900 GitHub stars on the open-source implementation.
- Designed the TARS framework (Target, Adoption, Retention, Satisfaction), now adopted across multiple product teams for feature evaluation and prioritisation.
Senior Product Manager, Knowledge Discovery
This was the bridge between Mendeley and everything that came after. I moved into shared services, building the search infrastructure that powered discovery across Scopus, ScienceDirect, Editorial Manager, and SSRN. The vector search platform I developed here became a 2023 strategic initiative and the technical backbone of what would eventually become Scopus AI.
- Built the Vector Search Platform that became a company-wide strategic initiative, deployed across Scopus AI and ScienceDirect.
- Created People Finder / Scopus Researcher Discovery, a cross-product service opening up a multi-million-pound revenue opportunity.
- Built Found Reviewer, improving peer review matching and increasing the 7-day invite-to-acceptance rate from 16% to 20%.
- Pioneered weekly gender representation measurement for reviewer selection, an Elsevier first that improved bias reranking models.
Product Manager
Celebrating with the Mendeley team
Coming in to improve acquisition and onboarding, Mendeley became my product management bootcamp. I learned how academics actually work: discovering papers, building libraries, writing research. Through proper A/B testing, user research, and cross-team collaboration, I implemented onboarding flows that actually worked. But the real education came from rebuilding search from scratch.
I led the team on the search rebuild, but I also used it as an excuse to get my hands dirty with the underlying systems: Elasticsearch, OpenSearch, lexical and boolean query building, graph search. I wanted to understand how retrieval actually worked, not just manage the roadmap for it. That grounding in search infrastructure meant I could lead strategy and product development with the team from a position of genuine understanding, not just stakeholder alignment. It's also the foundational knowledge I drew on years later when building the Scopus AI prototype and then LeapSpace.
- Rebuilt Mendeley's search from the ground up, creating a tool that genuinely helped researchers build better libraries.
- Doubled acquisitions after the search integration, a result that gave leadership confidence to trust me with bigger projects touching search across Scopus and other products.
- Improved onboarding flows that increased new user engagement 3x and overall product retention by 25%.
Team achievements and collaborative success at Elsevier.
Digital Director
At Medikidz, we made superhero comics explaining chemotherapy to kids, built games for clinical trial adherence, and were experimenting with augmented reality for insulin education before AR was something anyone talked about. Trying to make cancer understandable to a seven-year-old turns out to be excellent training for explaining anything complex to anyone, something I think about constantly when building AI tools for researchers.
This was also where I got my first real taste of building things to drive commercial goals and user value for a larger audience. I was prototyping games in Unity, building interactive comics in Canvas JS, and putting together the webshop and CMS for internal comic production. None of it was polished, but it gave me the technical vocabulary and instincts to work properly with engineering teams: not just managing a backlog, but being able to explore ideas hands-on and demonstrate them to senior stakeholders in a way that actually shaped product strategy.
In full superhero mode at a school event
Medikidz Explore: blood cell characters in the game I prototyped
- Our comics reached millions of children globally. Learning to make genuinely difficult things clear, without dumbing them down, has stayed with me more than almost anything else in my career.
- Grew user base by 200% while ensuring medical accuracy and child safety in everything we built.
Turning complex medical concepts into engaging stories for children through creative sketching and character development.
Medical Doctor
I started in Glasgow's NHS, rotating through general medicine, A&E, surgery, and obstetrics. Night shifts meant juggling five emergencies at once with multiple bleepers around my neck. It was intense, but it shaped how I think about everything: staying calm when things are moving fast, paying attention to what's actually in front of you, and never losing sight of who you're there for.
- Between shifts, I returned to programming for the first time since childhood, self-teaching Objective-C to build Ward Round, a medical education app for junior doctors on iPod Touch. That project was the start of everything that came after.
- Awarded Junior Doctor Representative of the Year by the Royal College of Physicians & Surgeons of Glasgow.
During my medical career at NHS Greater Glasgow & Clyde - the foundation that shaped my approach to problem-solving.
Junior Doctor Rep of the Year Award
Ward Round app prototype
Technical architecture planning
Side Projects
Fun experiments & weekend builds
IAM46: heart rate, location, and Instagram API hacking (2015)
Built a real-time tracking system streaming Apple Watch data to a custom dashboard. An early excuse to learn APIs, data visualisation, and PHP.
Postt: post on demand for London (2015)
A General Assembly capstone project. Pickup, package, and post on demand. Never wait in line at the post office again.
Paper of the Day Chrome Extension
Used ML to surface trending academic papers with beautiful Unsplash backgrounds. The idea was simple: make opening a new tab slightly more interesting.
3D Medical Education Characters
Designed 3D characters to teach children about insulin management. A leftover instinct from the Medikidz days.
Education & Development
University of Glasgow
Five-year medical degree combining clinical practice with evidence-based research training.
Mastering Product Management
Advanced product strategy, growth frameworks, and cross-functional leadership.
Product Management & AI/ML
iOS development (self-taught 2010-2013), product management fundamentals, and continuous AI/ML learning since 2018.
GA London, Autumn 2015
Patents & Innovations
RAG Fusion
A patent-pending approach to retrieval-augmented generation that improves how AI systems understand research queries beyond simple keyword matching. The technique has gained significant adoption in the research community, with the open-source implementation receiving over 900 stars.
SeqRAG
"Agents for the rest of us" started as a blog post about making AI agents less intimidating. Turned into a patent-pending framework for sequential reasoning. Sometimes the best innovations come from trying to explain things simply.
Continuity-Aware Temporal Clustering
A patent-pending method for clustering research content across time intervals with continuity awareness. Rather than treating each time window in isolation, the system tracks how themes persist, emerge, and evolve, letting researchers spot where a field is heading earlier in the cycle. This became the algorithmic foundation of the Emerging Themes feature in Scopus AI.
Speaking & Recognition
Selected Talks
ACRL Webinar: Empowering Research with GenAI and Agentic AI
1,861 registrants, 399 live attendees (August 2025)
UIDP Webinar: Rethinking Research in the Age of GenAI
168 registrants (August 2025)
NORDP Webinar: GenAI for Strategic Research Development
September 2025
Elsevier Product Excellence Academy: Vibe Coding Prototypes
Keynote, NPS 42 (February 2026)
NALAB: GenAI at North American Library Advisory Board
November 2024
Recognition
Media Spokesperson
Chronicle of Higher Education, Science Magazine, Axios, Vox Future Perfect, Space.com, Sueddeutsche Zeitung
Junior Doctor Representative of the Year
NHS Greater Glasgow & Clyde
Publications
Industry Publications
"Enhancing Search Relevance: User-Driven Personalized Vector Search"
Proceedings of the 7th Annual RELX Search Summit, 2024
"Using ChatGPT to Evaluate Search Results"
Proceedings of the 7th Annual RELX Search Summit, 2024
"An Efficient Method for Identifying Conflicts of Interest in Peer Review Using Graph Analysis"
Proceedings of the 6th Annual RELX Search Summit, 2022
Medical Publications
Get in Touch
If you want to talk about AI product strategy, research tools, or anything else on this site, I'm always happy to hear from people. LinkedIn is probably the easiest way to reach me, or GitHub if it's something technical, or Medium if you want to argue about one of my articles.
Looking for a speaker?
I speak at conferences about AI, product management, and why your RAG pipeline probably needs fixing. Also available for explaining complex tech through increasingly ridiculous analogies. Drop me a message on LinkedIn.
You scrolled all the way down! Yes, still a Dragon Ball fan.