In 2017, ECE Associate Professor Jeremy Holleman joined semiconductor veterans Kurt Busch, Pieter Vorenkamp, and Stephen Bailey, to form Syntiant, a startup company focused on ultra-low-power chips for machine learning. This October, Syntiant raised $25M in a Series B funding round led by M12, Microsoft’s venture fund.
Syntiant is developing semiconductor solutions designed from the ground up for deep learning inference. Without the constraints of legacy processor architectures, Syntiant’s neural decision processors (NDPs) use an analog neural network that can offer orders of magnitude lower power by extreme memory efficiency along with massively parallel computation with modest precision. Ideally suited for applications as small as hearing aids and IoT, or as large as smart speakers and mobile phones, Syntiant enables always-on deep learning inference in battery-powered devices.
Earlier this month at Infineon’s OktoberTech 2018, Syntiant demonstrated a prototype NDP that can simultaneously support dozens of application-defined audio and keyword classifications, enabling developers to create custom, always-on speech user interfaces. Optimized for audio data rates, the processor is capable of speaker identification, audio event detection, environment classification, and sensor analytics.
Dr. Jeremy Holleman is currently on leave from UNC Charlotte.