Algorithm Engineer for Camera Egomotion - Autonomous Mobility (TM)

Job Description

We are looking for a software engineer with a good knowledge of the C++ programming language and inclination towards mathematical models, mathematical optimization and image processing.

You will develop and refine (within a team) the algorithms that estimate the pose of the car and the cameras mounted on it as the car is moving, as well as the depth (3D position) of the surrounding scene based on inputs (e.g. optical flow) from moving cameras. You will use inputs such as optical flow, and/or estimated 3D positions to compute the position and orientation of the cameras until the requested positional and rotational accuracy are fulfilling the needs of downstream functionality (such as 3D reconstruction, pothole detection, rough road detection, free space detection, etc.), and meet the hardware constraints (most notably runtime).

#AutonomousMobility

• Ability to work with mathematical models and abstractions;
• Good knowledge and experience (3+ years) in C++. Experience in C, Rust, Java, C# or other multiparadigm language is also taken in consideration;
• Algorithmic thinking and familiarity with algorithms, especially those in image processing and signal processing (e.g. Kalman filters, correspondence and tracking, etc.);
• Able to handle design (based on principles), implementation and (unit) testing in an agile approach;
• Physics competencies (e.g.: motion models, coordinate systems, SI units, basic optics).

Can be a plus:

• Experience in embedded vision - egomotion, visual odometry and other areas related to camera geometry;
• Familiarity with camera models, projective geometry and different projection types;
• Probability theory (expected values, Bayes formula, maximum likelihood methods) and statistics (t-test);
• Python or MATLAB; experience in other prototyping languages (Julia, Wolfram, Octave, etc.) also taken in consideration;
• Familiarity with numerical methods (polynomial approximations, mathematical optimization, numerical analysis, IEEE 754, fixed precision and modern hardware floating point formats such as bf16 and tf32);
• Embedded systems know how (Compiler toolchains, CPU Architectures, debuggers and remote debugging, busses and communication protocols such as USB, Ethernet, TCP/IP and UART, hardware setup);
• Runtime optimization know-how (vector processing, memory access patterns, pointer aliasing, knowledge of perf profiler or other relevant profilers, CPU micro-architecture knowledge);
• Machine learning, including deep learning neural;
• Embedded machine learning (compilers, activation functions, quantization, sparsity, etc.);
• Linux (usual tools and shell scripts) and embedded Linux knowledge (flashing, booting and bootloaders, various kernel sub-systems such as GadgetFS or remoteproc);
• Familiarity with automotive standards (MISRA, AUTOSAR, ASPICE, etc.);
• Experience working with software requirements, requirements tracking tools and requirements traceability;
• Experience with formal verification (Ada/SPARK, Coq, Idris or others.

What we offer:

Pay for Performance:

• Achievement Bonuses and Rewards;
• Relocation Bonus for non-Timisoara Residents;
• Recommendation Bonuses for new team members;
• Flexibility Program including flexible hours, mobile work and sabbaticals.

Wellbeing:

• Health & Wellness (Private Health Insurance, Life Insurance, Sport activities etc.);
• Different discounts (glasses, tires, medical, shopping);
• In-house restaurant & coffee corners.

Life-Long Learning:

• Technical, Soft Skills & Leadership trainings;
• Dedicated Programs and Conferences;
• Free Language Courses (English, German, French etc);
• Access to e-learning platforms;
• Career development opportunities (local and international);
• Internal development communities (Experts, Agile Community of Practice, Artificial Intelligence etc).

Ready to drive with Continental? Take the first step and fill in the online application.

We develop solutions for Autonomous Mobility, an emerging and innovative field in the automotive business. Using cameras and lidar sensors, our end products are scanning the environment around the vehicle and provide active assistance, from making the driver aware of dangers, to braking or even steering the car out of the harm's way. The goal of this kind of systems is to help protect the lives of passengers and pedestrians alike.

In particular, the Detection Software group is handling various processing steps in the Computer Vision stack for driving scenarios (e.g. front looking cameras) and parking scenarios (e.g. surround vision cameras).