London Stock Exchange Group: £5M AI Programme Transformation

01 / Context

The challenge

The London Stock Exchange Group's Data Analytics division needed to modernize their development processes and leverage cutting-edge AI technologies to maintain competitive advantage in financial data services. The organization faced challenges with: • Manual, time-consuming development workflows across geographically distributed teams • Need for standardized AI/ML implementation across multiple product lines • Requirements to improve developer productivity while maintaining quality standards • Complex coordination across teams in 6 different countries and time zones

02 / Contribution

The work

As Program Manager, I led the creation and implementation of a comprehensive technical roadmap that transformed LSEG's approach to AI and development: **AI Integration Strategy:** • Implemented Generative AI tools using multiple LLMs (OpenAI, DeepSeek, Gemma 2, LLaMa) • Developed RAG (Retrieval Augmented Generation) systems for knowledge management • Created AI agents using Crew AI for automated workflow optimization • Built custom VS Code extensions and Excel add-ins for developer productivity **Programme Management:** • Managed geographically distributed team across 6 countries • Implemented Lean/Agile methodologies for rapid delivery • Established governance frameworks for AI implementation • Created technical roadmap aligned with business objectives **Technology Implementation:** • Python-based ML pipelines with Conda environment management • Integration with Hugging Face models and custom fine-tuning • Copilot and ChatGPT integration for enhanced developer workflows • Comprehensive monitoring and analytics dashboards

03 / Evidence

Reported results

18-20%

Improvement in development productivity through AI automation

£5M

Programme value delivered across multiple initiatives

25

Team members coordinated across 6 countries

Multiple

AI/LLM models successfully integrated and deployed

These results are reported in the existing portfolio and have not been independently verified.