Software Engineer - Video Pipeline

Irvine, CA /
Software – Perception /
Full-time
Anduril is a defense technology company, bringing Silicon Valley talent and funding to the defense sector. Our technology helps our customers solve their toughest challenges by enabling them to make better, more informed decisions in life-and-death situations. We’ve assembled a diverse team of experts in artificial intelligence, computer vision, sensor fusion, optics, and data analysis that are creating software and hardware solutions to radically evolve the capabilities of the United States and our allies. If you are passionate about solving problems that have real impact, come join Anduril and build the future of defense. 

At Anduril we are developing aerial, aquatic, and ground-based robotic systems. Lattice Ghost UAS, Sentry Tower and our other platforms stream video over unreliable networks from remote locations. The role of our video pipeline engineers is to develop real-time and asynchronous video feeds from remote deployments to web, virtual reality clients, computer vision front-end and back-end pipelines. Video pipeline engineers help design the ISR payloads for drones and the camera systems used in our other existing and future products. They also have some computer vision overlap with e.g. object detection. Extensive prior experience with relevant video technologies is not required for this role and is expected to be learned on the job.

Required Experience:

    • Embedded software experience with C++/Rust
    • BS degree in an engineering discipline
    • General understanding of camera hardware and video formats
    • Solid understanding of writing reliable, multi-threaded software

Desired Experience:

    • The following (below) is not required but does indicate a good fit for the role and will be learned on the job as needed
    • Embedded Linux software development experience
    • Tools: v4l2, ffmpeg, WebRTC SFU routers like Janus, OpenCV/Python
    • Formats: MP4, MPEG-TS, MISB KLV standards, H264 format, WebRTC/RTP
    • Image formats: Demosaicing raw images, YUV420, BGR
    • Linux platform: V4L2 camera API, VA-API, ffmpeg, gstreamer, OpenGL
    • Nvidia Jetson platform: Argus, V4L2 extensions, OpenGL extensions
    • Advanced camera controls: XU extension unit interface, region of interest (ROI), analog versus digital gain
    • Experience integrating cameras into computer vision pipelines
    • Image processing experience with: Denoising, contrast enhancement, super-resolution, auto-focus, auto-exposure, auto-white-balance
    • MIPI camera bringup (Linux kernel drivers)
U.S. Person status is required as this position needs to access export controlled data.