Job Summary
Own the camera software stack on ARM-based SoC platforms — ISP pipeline configuration and tuning, sensor driver bring-up, and multi-camera synchronization for vision-heavy industrial, robotics and enterprise edge products.
Key Responsibilities
- Bring up camera sensors and ISP pipelines end to end — sensor drivers, MIPI CSI-2 configuration, V4L2/media controller graphs and HAL integration.
- Configure, validate and support tuning of ISP blocks: 3A (AE/AWB/AF), lens shading, demosaic, noise reduction, HDR, tone mapping and colour pipeline.
- Enable multi-camera use cases — hardware and software synchronization, frame alignment, multi-stream concurrency and bandwidth/latency budgeting.
- Integrate camera output with downstream vision, encode and AI inference pipelines with zero-copy buffer sharing where possible.
- Characterize and improve image quality; work with tuning tools and objective/subjective IQ evaluation methods.
- Debug frame drops, timing, buffer starvation and image artefacts using trace, ISP register dumps and capture analysis.
- Support new camera module qualification and work with hardware and module vendors on interface issues.
- Define camera validation criteria and contribute to automated camera regression coverage.
Skill Requirements
- 8–10 years in embedded camera / imaging software, including 4+ years on ISP-based platforms.
- Strong C/C++ on embedded Linux and/or Android.
- Deep V4L2, media controller framework and MIPI CSI-2 knowledge; camera sensor driver development.
- Practical ISP pipeline understanding — 3A behaviour, sensor modes, exposure/gain control, RAW to YUV processing.
- Multi-camera system experience including synchronization and concurrency management.
- Buffer management and memory sharing frameworks (dma-buf, ION/DMA heaps, GBM).
- Systematic debugging of real-time streaming defects.
Other Requirements
- Android Camera HAL3 / CameraX / Camera2 stack experience.
- Image quality tuning experience with vendor tuning tools; colour science exposure.
- Depth, stereo, ToF or event-camera pipelines.
- GPU or NPU based pre/post-processing of camera frames.
- OpenCV, libcamera or GStreamer camera source integration.