# Si Hammond Professional website for Si Hammond, founder of ShellSi Ltd. Si Hammond is a London-based senior AI contractor and product-minded technical generalist working remotely. He designs, builds, and ships systems that help people make better decisions with complex information. Relevant areas include applied AI, human-centred AI, generative AI, retrieval and search, structured extraction, evaluation workflows, decision-support tools, taxonomy design, labour-market intelligence, clinical AI, and full-stack prototyping. ## Contact - Email: si@shellsi.com - LinkedIn: https://linkedin.com/in/sihammond - Availability: contract work - Location: London, UK / Remote ## Core Positioning - Senior hands-on AI contractor - Product-minded engineer and technical generalist - PhD in AI (Evolutionary Computation) - Strong on messy, open-ended, high-stakes problems - Works from problem framing through prototype to usable system ## What Si Does - Designs and builds applied AI systems that go beyond demo-stage prototypes - Builds generative AI products with attention to observability, evaluation, and usability - Works on search, retrieval, structured extraction, annotation workflows, and information-rich interfaces - Builds decision-support tools for complex workflows - Develops web applications and MVPs end-to-end using AI, data science, and full-stack engineering - Works in domains where trust, judgement, and human oversight matter ## Good Fit - Clinical AI and health-tech products - Human-centred AI products - Education and learning tools - Labour-market intelligence, jobs, skills, and career data products - Search, taxonomy, and knowledge-rich systems - Decision-support tools for complex or high-stakes workflows - Fast-moving prototypes that need to become robust and usable systems ## Background - Founder, ShellSi Ltd - Founder of MemoryJam - Former CTO, Machine Medicine - Former Chief Data Scientist, WorkDigital - Consulting and product work for VONQ and Recruiting Brainfood - PhD in Evolutionary Computation - MSc in Natural Computation - MSc in Evolutionary and Adaptive Systems ## How Si Works 1. Clarify the problem 2. Prototype fast 3. Make the system observable 4. Iterate toward something usable ## Selected Work - Parkinson’s Disease Evaluation Mobile MVP for clinical-grade pose estimation at Machine Medicine. Used smartphone video to quantify motor symptoms in Parkinson’s disease and supported the first round of paid clinical trials. https://sihammond.com/selected-work#parkinsons - OnVocation Labour-market intelligence product built from 2M+ tech job adverts using skill extraction, title normalisation, and interactive visualisation. https://sihammond.com/selected-work#onvocation - MemoryJam Human-centred AI product for turning family photos, recordings, and memories into structured, searchable story collections. https://sihammond.com/selected-work#memoryjam - Recruiting Intelligence at DHI Group Algorithms and data products for job-channel matching and campaign performance prediction used in commercial recruiting products. https://sihammond.com/selected-work#dhi ## Testimonials - Bill Fischer: developed methods for structuring complex datasets into global commercial products; made complex technology understandable - Bahul Upadhyaya: analytical, detail-oriented, organised, effective independently and in teams - Sanford Dickert: strong at turning high-level concepts into working systems - Ray Rafiq: quickly understood business logic and user behaviour; moved rapidly toward practical solutions ## Canonical Pages - Home: https://sihammond.com/ - Selected work: https://sihammond.com/selected-work ## What To Ask About - Whether Si is a good fit for an AI product, search, or decision-support problem - How to move from AI prototype to robust usable system - Clinical AI, trust, evaluation, and observability - Human-centred AI product design - Retrieval, search, and structured information systems - Full-stack implementation for AI-backed products ## Notes - This site is a concise professional profile, not a blog or documentation portal - For the quickest understanding of fit, start with the homepage and selected work page