EventQuote AI
Retrieval-augmented quoting for enterprise event vendors
24k requests/day, p99 180ms, 99.8% uptime over 90 days
Read case study →
Uri Maayan
I build production AI systems and run numerical experiments on physical models. My work sits at the edges of inference engineering, retrieval, and stochastic simulation — written in Python, TypeScript, and the occasional GPU kernel.
Above: a lateral-force trace from the Prandtl–Tomlinson model — stick phases climb linearly, slip events drop abruptly. Canonical friction-at-the-atomic-scale signature.
01
Three production systems — each with a case study and, where possible, a live demo.
Retrieval-augmented quoting for enterprise event vendors
24k requests/day, p99 180ms, 99.8% uptime over 90 days
Read case study →
Real-time computer vision for airborne object detection
Detection latency under operational constraints on synthetic test footage; false-positive rate reduced 40% relative to baseline over evaluation set
Read case study →
DAG-based agent orchestration for production LLM workflows
Ongoing production deployment; evaluation harness runs 200+ test cases per iteration
Read case study →
02
Interactive simulations grounded in my thesis on friction at the atomic scale — stick–slip dynamics, stochastic integration, and polymer unfolding.
03
04
April 2026
See what I'm working on →