Dejan Pangercic

CTO & Co-Founder · Apex.AI

Autonomous systems software — safety-certified middleware and runtimes for embedded AIPalo Alto, California

Dejan Pangercic is co-founder and Chief Technology Officer of Apex.AI, the Palo Alto company building safety-certified software for autonomous and embedded systems. He leads development of its product line — Apex.Grace, the deterministic execution runtime formerly named Apex.OS, alongside the Apex.Ida middleware and the Apex.Alan deployment tooling — and came to it from a robotics research career spanning the Technical University of Munich, Bosch's Palo Alto research centre, Bosch's Deepfield Robotics venture and Faraday Future. He is a longstanding figure in the open-source robotics community, an early ROS developer, a Point Cloud Library contributor and a member of the ROS 2 Technical Steering Committee.

Last reviewed: Sep 19, 2026

Dejan Pangercic

CTO & Co-Founder Apex.AI

Last Reviewed: September 19, 2026

Executive Summary

Dejan Pangercic co-founded Apex.AI in 2017 to address a gap that anyone who has tried to ship an autonomous system runs into. The open-source robotics ecosystem — ROS above all — gave researchers an enormously productive way to build perception and autonomy software, but ROS was never designed to be certifiable, deterministic or safe in the sense that a vehicle regulator or a medical-device authority means by those words. The result was that a generation of robotics teams prototyped in ROS and then discovered they could not ship it. Apex.AI’s answer was to rebuild the ideas with automotive-grade rigour: Apex.Grace, launched in 2020 as Apex.OS, provides deterministic execution of embedded software; Apex.Ida provides fault-tolerant middleware for moving data between components; Apex.Alan handles safe deployment and CI/CD. Pangercic leads that development as chief technology officer, and his credibility for the task is unusual — he is an early ROS developer and a member of the ROS 2 Technical Steering Committee, which is to say he helps steer the open-source project his company’s products are designed to make shippable. The company has raised more than $75 million from investors including Continental, Daimler Truck, Volvo Group, Toyota Ventures, Airbus Ventures and AGCO, and now sells into defence, automotive, agriculture, medical and AI-native systems.

Career Highlights

  • Co-founded Apex.AI with chief executive Jan Becker and Tavis Szeto, and serves as chief technology officer, coordinating all development activity.
  • Led development of the company’s first product, launched in 2020 after three years of work as Apex.OS and since renamed Apex.Grace, a deterministic execution environment for embedded software.
  • Extended the line to Apex.Ida middleware, Apex.Alan deployment and CI/CD tooling, Apex.OS for Verification and Validation, and Apex.Autonomy for unmanned ground vehicles.
  • Apex.AI has raised more than $75 million from investors including AGCO, Airbus Ventures, Continental AG, Daimler Truck, Toyota Ventures and Volvo Group, and operates from Palo Alto with offices in Berlin, Munich, Stuttgart, Tokyo and Pangyo.
  • Member of the ROS 2 Technical Steering Committee and an early developer of ROS.
  • Contributor to the Point Cloud Library, the open-source standard for 3D perception.
  • Author of several papers in robotics and robotic perception, holder of two patent applications in field robotics, workshop organiser and regular reviewer for leading robotics conferences.
  • As chief technology officer of Bosch’s Deepfield Robotics, built an autonomous asparagus-harvesting robot and a 4D scanning robot for plant emergence counting and 3D leaf-area measurement.
  • Led perception and framework engineering at Faraday Future before founding Apex.AI.

Professional Journey

Pangercic trained in electrical engineering at the University of Ljubljana, taking both a bachelor’s and a master’s degree there, before moving to Germany for doctoral work. His research assistantship at the Technical University of Munich and his doctorate in computer science from the University of Bremen place him squarely in the German robotics research lineage that produced a great deal of the early ROS ecosystem — the same community in which he became a contributor to the Point Cloud Library and an early ROS developer. That open-source standing is not incidental to his later career; it is the foundation of it.

He moved from academia into corporate research as a senior research engineer at the Bosch Research and Technology Center in Palo Alto, and then into applied robotics as chief technology officer of Deepfield Robotics, Bosch’s agricultural robotics venture. The Deepfield work is more interesting than it sounds: an autonomous asparagus harvester has to perceive an irregular, partly occluded organic object in an unstructured outdoor environment and act on it with millimetre precision, and a 4D scanning robot measuring plant emergence and leaf area has to do so repeatably across a growing season. These are hard perception problems with no tolerance for the sort of best-effort behaviour that research code gets away with.

He then joined Faraday Future as perception and framework lead, working on the perception stack and the software framework beneath it for an electric-vehicle programme. The pairing of “perception” and “framework” in a single remit is the clue to what came next: having built perception systems at Bosch, at Deepfield and at Faraday Future, he had repeatedly confronted the absence of a production-grade framework to run them on.

In 2017 he co-founded Apex.AI with Jan Becker, who had also come out of Bosch’s autonomous driving research, and Tavis Szeto. The founding proposition was to take the developer productivity of ROS and rebuild it to automotive functional-safety standards, so that software prototyped in the open-source world could be certified and shipped. The first product took three years and launched in 2020 as Apex.OS, later renamed Apex.Grace.

The company’s trajectory since has been one of widening rather than pivoting. Apex.Ida added fault-tolerant middleware for data flow; Apex.Alan added automated tooling and CI/CD workflows for safe deployment; Apex.OS for Verification and Validation added data recording and replay for testing; and Apex.Autonomy added a full autonomy stack for unmanned ground vehicles. The customer and industry mix has widened correspondingly, from automotive into defence systems, agriculture, medical devices and what the company now calls AI-native systems, with cited users including Brunswick, MOIA, AGCO, KRONE and LEMKEN, and Tier IV. The investor base reflects the same spread — Continental, Daimler Truck and Volvo Group from automotive and trucking, AGCO from agriculture, Airbus Ventures from aerospace, Toyota Ventures from mobility — which is the pattern of a company whose product is infrastructure for other people’s regulated hardware.

Throughout, Pangercic has kept his position in the open-source community rather than retreating from it, sitting on the ROS 2 Technical Steering Committee while running the engineering organisation of a company that commercialises safety-certified alternatives to it.

Education

University of Bremen

Doctor of Philosophy, Computer Science

University of Ljubljana

Master’s degree, Electrical Engineering

University of Ljubljana

Bachelor’s degree, Electrical Engineering