معرفی
Vladimir Loncar is a researcher specializing in machine learning, FPGA optimization, and high-energy physics computing. His work focuses on accelerating neural networks and scientific algorithms using hardware-aware techniques. Notably, he contributes to the hls4ml framework for FPGA deployment of machine learning models, and has applied these methods to particle physics experiments like LHCb and the HL-LHC. His research spans symbolic regression, recurrent neural networks, and real-time data processing for large-scale physics detectors.
Key projects include developing resource-efficient inference systems (e.g., Tailor for CNN optimization), benchmarking frameworks for GNN-based surrogate models, and latency-critical implementations for collider experiments. Loncar's work bridges theoretical physics and computational engineering, emphasizing practical applications in experimental particle physics, quantum simulations, and autonomous detector control. He collaborates extensively with institutions like CERN and the sPHENIX collaboration.
