I am one of the first Agentic AI specialists, having committed to the technology in 2023, as soon as it was feasible. Project examples:- Agents to Fully Automate Jupyter Notebooks for Data ScienceBuilt AI agents enabled to solve data science projects. Agents access data, discuss plans, execute code, and correct themselves. All in Jupyter notebooks for data science. See video of fully automated solution in action: https://youtu.be/iF2RguoqmrA. This was the first stateful code executor for Microsoft Research's Autogen project. Became Top 3 most popular projects for AutoGen and was incorporated into the product by Microsoft Research.- Agents for Knowledge Graph ConstructionDeveloped a pipeline which converts unstructured text into a searchable graph using LLM's to identify triples (subject object predicates) for a knowledge graph. Searchability was enabled via RAG and the WordNet hierarchy. Negatives in predicates were examined using a multiNLI model- Agents for Market Research & Product DevelopmentBuilt AI agent solutions for clients seeking to automate marketing processes:- Online research with reliable sources and fact checking- Web search thousands of contacts to identify which were actively posting news which suited client services, then filtering those leads in a structured qualification pipeline- Conduct sustainability market research interviews as conversations, then construct recommnedations reports for those interviewees- Draft user requirements in AzureDevOps format from transcriptions of software user conversations.

Consulted with R&D clients to outline programs of data science strategy, quantify business value, review technologies, manage teams and oversee agile delivery of data science and ML solutions. I also brought my renewables expertise to highlight new sustainability opportunities. Examples of engagements:• Engineered a strategy for global R&D teams to collaborate on machine learning studio and share standardised data, using Microsoft AzureML drawing from a managed datalake. Client was in chemicals and sought to grow solutions for electric vehicles.• Counseled large renewable construction company (solar and hydro) on managing a data governance project and demonstrated how a datalake is iteratively justified via a Digital Twin strategy• Consulted on MLOps roles and data science team structure for top tier automotive manufacturer• Identified and developed opportunity with Natural History Museum to deliver global data on the sensitivity of species diversity to global warming, timed for COP27 at Sharm El Sheikh.These projects were aided by my work to distil and productise Capgemini's R&D experience in AI for client proposals and tenders:• Capgemini Engineering's Sustainability Offer• How to Manage Data Science Programmes at Scale• Edge Computing + Privacy; Encryption technologies and Federated Learning for automotive• Data Driven R&D for Chemicals Industry (data governance and strategy)

Using data, analytics and machine learning to automate processes, manage finance and enhance the relationships of this agricultural machinery business. Grew the turnover of this family firm from £18m to £35m. Doubled shareholder wealth through modernisation and professionalisation of services.• Led a team to implement an inventory optimization tool which I prototyped using python and time series models inspired by the deep learning architecture, WaveNet. This replaced my own previous ARIMA models and regression tools built in R• Designed and financed renewables investments to reduce carbon footprint of company. 300+ solar modules, designed in PVsyst. Cladding of buildings for insulation. Lighting replacement programme. Organised DNO work to bury 11kV lines, replace switchgear and transformers etc.• Developed a fraud detection tool to meet Financial Conduct Authority requirements. This followed exploration of numerous deep learning architectures, autoencoders and exploration of various GAN architectures. I joined the team to develop these tools via Azure Machine Learning, Visual Studio Code and Python on Ubuntu• Set strategy for company finance and analytics, managing an accounting team in addition to analytics• Worked with ERP provider to implement improvements for data quality; e-billing, paperless office, bar-coding, warehousing tasks via mobile etc• Developed a Business Intelligence system in SQL Server, SSRS/SSIS and C# to automate and personalize complex reporting tasks such as credit risk analysis, inventory valuations, monthly P&L etc. Externally audited• Researched generative deep learning models in python/tensorflow for business planning, using automated downloads of all company reports at Companies House (UK) and the Securities & Exchange Commission (US).

My first brief with this business was to modernise it. This became a programme of works in software, networks, construction and team building.• Migrated IT to cloud (Azure Active Directory Premium / Sharepoint ) for this multisite business, which greatly enhanced security and eased system management. Then outsourced the WAN and IT support tasks • Personally project managed the design, finance and construction of two new warehouses to CDM regs. • Devised and implemented processes to achieve Financial Conduct Authority authorization as a consumer credit organisation (FRN677769), including the establishment of a new credit control team• Persuaded key staff to form an MBO team for part of the trading operations, providing succession planning and incentivising a team of multi-disciplinary, multi-location managers • Established the need and implemented a plan for restructure of the business: Demerger into holding and trading co, new staff share scheme, closure of loss making subsidiary, share buybacks

Board responsibility for Product and Software Development This company of 50 staff is now a part of the data and analytics business, Cirium.• Instructed project managers throughout a multi year programme of software product development. Multiple full lifecycle projects, £1m budget per yr. Proved market demand before investing• Worked in multi year partnership with actuary to develop analytics required for insurance and banking industry. Directly managed lead tester and lead developer, applying agile software development (DSDM), within corporate (Prince 2) projects to adhere to business ISO certification.• Developed consultancy products as well as software. Opened a new business line for Carbon management including regression modelling of network fuel demand and business cycle forecasting.• Retained Lloyds Development Capital as private equity investor having identified opportunity in the competitor space. Agreed 2yr strategy and £2m budget for product development• Pioneered cloud technology with use of Endeca search and Akamai CDN, greatly increasing the global customer base as we transitioned from on-premises software to SaaS• Managed the programme and product roadmaps, establishing a user committee keen to beta test products and become launch customers in Asia, US and Europe

• Deployed infrastructure for the Hyperion accounting system across Ford's European offices under the compliance constraints of this bank, fully regulated by the FSA. The objective was securitisation of vehicle loans.

Global Freight Exchange was a dot com start-up initiated by McKinsey employees and funded by Kerry Packer, Lufthansa, British Airways and American Airlines, amongst others. Acquired by Descartes, the Canadian logistics and data company.• Developed first online marketplace for air cargo at Global Freight Exchange through business process redesign with flag carriers inc BA, AA, Lufthansa and freight forwarders inc Panalpina, Kuehne Nagel, Swiss etc. My role was in interviewing users and flow charting the freight booking process for the software developments requirements (RUP, UML).• Developed one of the first neural networks for sales management (2001)

Air Foyle specialised in the An124, the world's largest air freighters, for emergency relief, supporting the RAF and flying outsize industrial machinery around the world. It also operated B737's on behalf of easyJet and BAe146 for TNT• Developed software tools for freight network optimisation, maximising revenue tonne kilometres.• Gravity regression modelling of freight demand and network optimisation for multiple airline business plans• Lead analyst on a Public Private Partnership project to provide freight airlift for the UK Ministry of Defence.