Tomer Ast is an engineering leader building AI assistants that understand how people work: their relationships, routines, commitments, and changing priorities. His work on monday sidekick centers on turning scattered workplace activity into actionable context.
Ast studied at Tel Aviv University and worked with DealSize.ai on natural-language generation for financial content. He also helped develop FacetAI with Erez and Oren Braverman, creating realistic jewelry visualizations from customers’ ideas and sketches. His FacetAI-associated Hugging Face account includes a publicly listed model.
An engineering manager affiliated with monday.com at AI Engineer World’s Fair 2026, Ast collaborated with Omri Bruchim on the Monday world model, which organizes work around three elements: relationships between people and tasks, immediate operational signals, and durable patterns in individual behavior.
- Slow and fast context engines: A slower system identifies recurring routines, collaborators, goals, and working patterns; a faster system tracks overdue tasks, recent commitments, and emerging priorities. Their combined output gives an assistant both historical context and situational awareness.
- Precomputed work context: Workplace relationships and behavioral patterns are assembled before requests arrive, allowing the assistant to reason from an established representation instead of reconstructing priorities from disconnected records.
- Resilient context serving: Isolated data sources, live verification, and previously validated context prevent unreliable inputs from collapsing the assistant’s understanding.
Ast also acknowledges unresolved limitations: new users have little behavioral history, inferred patterns can reproduce bias, and representations of work inevitably lag reality. His conference presentation with Bruchim frames the practical challenge as distinguishing genuinely important changes from everyday noise.
