Crowdsourcing Industrial AI at National Scale: Lessons from DataRace

How does a country solve its toughest industrial problems when talent is scattered and data is messy? Azerbaijan's answer is DataRace — the National Artificial Intelligence Center's open-competition platform that turns real operational challenges into deployable AI, built by a nationwide community of engineers and data scientists.

In this session Mr.Ulfan will show how DataRace works as a repeatable engine: an industry or infrastructure operator brings a concrete problem — predictive maintenance, energy-demand forecasting, leak detection, logistics optimization — NAIC frames it as a competition with real data and clear success metrics, and hundreds of participants compete to produce working models at a fraction of the cost and time of traditional procurement.

He'll walk through what makes these competitions deliver in critical-infrastructure settings across energy, water, oil & gas, and transport: how to prepare and clean industrial data, guard against leaderboard overfitting, validate models for real deployment, and move a winning solution from notebook to production — along with the pitfalls, from OT security and data sensitivity to the hand-off from competitors to operators.

Attendees will leave with a practical blueprint for using open innovation to accelerate their own industrial AI: how to scope a problem, structure a challenge, and build the talent pipeline that sustains it. DataRace shows that a nation — or a company — doesn't need to build every model in-house. It needs the right problem, the right data, and a community ready to solve it.