R&D and AI
Certification in ‘Machine Learning, Applications for Business’ - LSE (April 2021) to gain a deeper understanding of managing data, supervised and un-supervised learning. Fundamentals of AI and smart technology.
‘ISO/IEC 42001:2023 Artificial Intelligence Management System Awareness’ - BSI (January 2025) in order to support AI teams and organisations as providers or developers of AI Systems.
Compliance: ISO/IEC 42001 AIMS governance framework and implementation requirements. The review of the EU AI Act extracting requirements for AI System Providers and Deployers. Defining the difference between an AI System and traditional machine learning applications. Governance, security, data management and standards accreditation (CE Marking).
AI Product and Projects Management
AI Product Management and Notion. While there are some similarities to the standard SDLC (Software Development Life Cycle) there are several obvious differences …for instance the understanding that the UX path a user shall follow, could be based on a series of semantic steps that need to be controlled if the user is meant to complete a specific task (user story). The final output from the model may be generated in Markdown format, that requires another step to be re-rendered to appear on the frontend UI. And triggering of model API’s can be costly which will require variants depending on the specific user story.
Notion, while not yet a replacement for JIRA, does feel like a platform that was developed with AI thinking at its core (instead of a bolt-on). Building flexible workspaces, databases and connected wiki’s that you can ‘ask’ - engage with intuitively. This is the direction of travel of PM tools, for sure!
Supporting the founders as AI Product Manager of enginuityai.io within a very small start-up team of specialists, building an AI product for research and innovation. Implementing RAG and AI Agents the eventual system shall be provisioned on MS Azure utilizing the GPT API for its models. Establishing project structure in Notion, Technical Documentation and Risk Management (based on ISO/IEC 23894) compliance, building the Product Development Plan of Record - an MVP pipeline is underway following a successful beta launch towards the end of 2024.
A project involving the strategy, research and implementation of various generative AI tools, principally LLM AI Assistants - for the purposes of training and education. Covering infrastructure evaluation, workflow integration and maintenance (accuracy and security). Policy writing : gathering user insights and advising stakeholders. Security over data sources and documentation storage, access controls with engagement safeguarding and protection over organisation IP. What does ‘trust’ mean in this space? Comparing the integration of either turnkey solutions vs custom built clients comprising third party APIs, cloud services and technologies such as RAG (Retrievable-Augmentation Generation). In this rapidly evolving landscape, is there a future-proof approach?
Machine Learning initiative for Google’s AI Impact Challenge - working alongside a creative technologist / data scientist to develop a programme of work for a machine learning project - defining the problem and identify the models / algorithms that will help find our solution. Project planning and scoping. Using the basic but very effective Tree Search model see here a chart to test if your solution ideas pass the ‘but is it using AI’ test.
Toaster R&D
Spanning various R&D initiatives which included AR, WebVR, gamification and pushing the limits on PWA (Progressive Web Apps) - the full-stack development team at Toaster were true magicians.
Create something