Tom Hubregtsen

Building electrified autonomous vehicles for farming and defense

Extropic, formerly known as Qyber, embeds generative AI into physics-based computers, achieving the limits of energy efficiency. Limited public disclosure, info on the $14.1M fundraise available at the attached links.

Simulated quantum machine learning circuits on Google's largest TPU cluster, to benchmark quantum computers. Work performed at X, the Moonshot Factory, formerly known as Google X.

Defined the BMW quantum roadmap. Aligned it with German & European initiatives, and oversaw internal quantum projects to ensure roadmap execution.

Performed research on Quantum accelerated Machine Learning algorithms, published at IEEE and ACM conferences and leveraged towards PhD degree.

Designed, lead and executed multiple projects to implement the BMW Group AI roadmap, ranging from proof of concepts to in-vehicle deployments to customers in Europe & China. 10+ patents pending, papers published in IEEE and ACM proceedings.

Performed research and technical management of projects in the fields of Big Data and ML, with a focus on neural networks, natural language processing for chatbots and distributed computing. Technical Sales as part of IBM Global Services team to board members of Forbes-100 companies in Europe, US and Asia-Pacific. Several patents files and papers published.

Performed IBM's first research into Apache Spark and Databricks. Research used for technical validation as part of IBM's $300M investment in Databricks and the creation of the Spark Technology Center in San Francisco, to accelerate AI and Big Data usecases.

As part of my thesis completed at IBM Research - Austin, I investigated whether user-addressable flash could close a performance gap for Apache Spark between data objects (Spark Resilient Distributed Data) stored in memory and such objects written out or spilled to a file system, to accelerate Big Data and AI.

Research in HPC datacenter topologies, including setting up a multidimensional (1-6) Torus topology simulation that has been verified to hold up to 50k nodes. This Torus topology is implemented in top-ranking supercomputers, such as IBM’s Blue Gene/Q and Fujitsu’s K Computer.

Thesis submitted and currently under review. Topic: Quantum accelerated Machine Learning. Supervisor: Prof. Jens Eisert.

Graduated in Jan 2015. Thesis topic: Evaluation of different storage systems for Apache Spark and Apache Hadoop Supervisor: Peter Hofstee (IBM Research)