AI Embedded Engineer (d/f/m)

Gültig bis

30.03.2023

Branche

Sonstige Berufe, Fahrzeugbau und Zulieferer, Technische Berufe und Ingenieurwesen, Internet, Web- un

Ort

Lübeck

URL

Zum Anbieter

Kontakt

Telefon

E-Mail

Beschreibung

Your new work environment
Welcome to gestigon - a Valeo brand. As one of the leading companies worldwide Valeo develops, manufactures and distributes components, integrated systems and modules for automobiles and utility cars. As part of Valeo’s Comfort & Driving Assistance Systems Business Group, the site in Lübeck specializes in software development, machine learning and artificial intelligence solutions. It offers sensor-agnostic disruptive software for our Interior Cocoon solutions, in particular gesture recognition and behavior tracking. The office and lab of gestigon are located in Lübeck, Germany.

Your Mission
Together with your team you develop solutions ranging from comfort/entertainment to safety-critical in an automotive environment by using cutting-edge algorithms. You are responsible to efficiently run real-time AI components to solve radar, computer vision or image processing problems on a given embedded target. You implement relevant optimization algorithms to reduce model runtime as well as its memory footprint. You know how to benefit from the inference engine specifics and hardware accelerators. You belong to a development team and are mainly at the interface with the function owner, the design and validation teams. Ideally you are knowledgeable in the algorithm domain as well and are able to influence the architecture so that it is better adapted to the target resources and constraints.

Your Responsibilities

• Work together with AI engineers to embed components ensur-ing optimal use of the targeted hardware
• Apply relevant optimization techniques (accelerated architec-ture, quantization etc.) to fulfil targeted hardware resources
• Precise evaluation of the impact of the implemented techniques on application KPI
• Run tests to provide estimation of runtime and resources use and track trade-off between accuracy vs optimization