Supporting UK Sports Institute’s Transition to New AI and Data Strategies – from Human Resources to Athlete Support

Nebuli - UK Sports Institute’s Private Generative AI Connectivity with Nano AI Workspace

The UK Sports Institute (UKSI) deployed Nebuli’s Nano Workspace generative-AI ecosystem and AIQ language models to support their teams in becoming more data-driven – from human resources to athlete support.

Key Services

Specialist Large Language Models
Private Generative AI
Data Strategy
Data Pre-processing
Data Modernisation
Smart Search

Key Markets

Academic Research
Human Resources
Sports Performance Data

Objectives

The UKSI provides support services to British Olympic and Paralympic sports, enabling sports and athletes to realise potential and achieve excellence.

The project aimed to support UKSI’s mission of becoming more AI-driven and empowering individuals with Nano’s conversational interface for an easier, nontechnical approach to accessing data insights.

We focused on unifying data ecosystems and applying AI models at scale through Nano AI Workspace, built on private dedicated cloud infrastructure.

Nano Private AI Workspace Ecosystem – Nebuli

Nebuli’s Solution

The key challenge we needed to prioretise was the necessity for wider engagements and connectivity among different stakeholders, with a wide variety of technical skills, to accomplish the data unification ambition.

The project also addressed common concerns related to using public Large Language Models (LLMs), such as security, relevancy, long-term lock-ins and costs.

Nebuli’s Nano Workspace offered the solution to address these critical points, including deploying multiple LLM models with a private and secured cloud server architecture.

The key elements included the following:

Data preprocessing and convergence of UKSI’s internal documents from multiple departments, such as HR.
Data segmentation, with a particular focus on identifying communities within the same database.
Applying Nebuli’s smart search API for full-text and abstracts across similar documents, generating a unique measure of similarity.
Private Vector Database of document classification classifies, ensuring documents are automatically assigned to the correct category and embedded into AIQ framework as the knowledge source of chosen LLM models. This removes the burden of manually searching through thousands of documents on hundreds of categories within shared storage servers.

Outcomes

With Nebuli’s Nano AI Workspace, UKSI teams and departments, including nontechnical staff members and athletes, do not need to worry about any additional technical setup or development of additional AI layers for the system to work.

Furthermore, this project has fulfilled the desire among UKSI teams to learn and test the power of generative AI models in their daily tasks in a truly private, sovereign, safe and highly personalised workspace. It has broken some of the key barriers to entry into AI work for nontechnical staff members.

Innovate with us

Nano AI Workspace

Get Started