28 GlowByte Teams Put Generative AI to the Test on a Fictional Retailer
GlowByte held an internal AI hackathon called AI Seasons, with the strongest engineering solutions developed by participants going on to form the basis of the company's internal development standards.
The four-week marathon brought together 28 teams and offered a prize pool of 275,000 rubles (about $3,500). All participants tackled the same challenge: build a data warehouse and BI analytics system for a fictional retail chain, MegaByte, with hundreds of stores and millions of orders, using generative AI at every stage, from data models to dashboards.
Projects were evaluated on three criteria: prompt engineering, technical implementation and business value. The winners took their projects to a customer-ready stage, with deployable demo environments, multilayered data-quality controls and documentation for on-call engineers.
Following the hackathon, most of the winning teams' solutions and procedures have already been incorporated into GlowByte's internal standards and are now being used in client projects. Templates for deploying demo environments, data-quality control procedures and practices for documenting AI development are particularly valuable for the business.
