Software Engineer (Embedded Systems Firmware)

📋 Description

• • Own the end-to-end optimization of operating systems that power body-motion video games running on mobile devices, smart TVs, and emerging form factors—your code will be the invisible force that lets millions of players jump, punch, and dance without lag or dropped frames.
• • Deeply integrate software and hardware to unlock new system capabilities: you will expose NPU, ISP, and GPU pipelines to higher-level applications, enabling real-time pose estimation, gesture recognition, and immersive AR overlays that were previously impossible on consumer-grade hardware.
• • Architect and maintain custom Android and Linux distributions, stripping away bloat while adding deterministic scheduling, low-latency camera paths, and thermal-aware throttling so that frame-perfect gameplay survives marathon sessions in Hong Kong’s humid summers.
• • Profile and shave milliseconds off critical paths using perfetto, systrace, Arm Mobile Studio, and custom tracing hooks; your weekly reports will quantify improvements in CPU utilization, memory bandwidth, and battery life that translate directly to five-star user reviews.
• • Build and customize AOSP from the ground up: fork HALs, craft Board Support Packages, and maintain a bulletproof CI pipeline that turns source code into signed firmware images in under 15 minutes—every commit you merge ships to production within days.
• • Design bootloader and secure-boot flows that protect user data while still allowing rapid OTA updates; you will sign images, manage key rotation, and coordinate with hardware vendors to ensure zero-downtime rollbacks if anything goes wrong.
• • Mentor a cross-functional team of Android app developers, ML engineers, and QA testers; run weekly deep-dives on cache coherency, memory barriers, and SIMD intrinsics so that high-level features are conceived with low-level constraints in mind.
• • Collaborate with camera-tuning experts to squeeze every photon out of ISP pipelines: tune noise-reduction kernels, craft HDR fusion algorithms, and synchronize multi-camera arrays so that body-tracking accuracy remains sub-millimeter even in dim living rooms.
• • Experiment with bleeding-edge neural-network accelerators; port TensorFlow Lite delegates to proprietary NPUs, profile layer-by-layer latency, and craft quantization schemes that shrink models by 10× without sacrificing precision.
• • Champion a culture of rigorous benchmarking: maintain a lab rack of reference devices, automate performance regression tests, and publish internal dashboards that celebrate every 1 % improvement in frame time or 1 mW reduction in power draw.
• • Influence product roadmaps by translating technical trade-offs into business impact; your “what-if” spreadsheets will help executives decide whether to invest in faster DDR or a larger NPU, ensuring engineering bets align with revenue targets.
• • Stay plugged into the global Arm ecosystem: attend Linaro Connect, contribute patches upstream, and maintain relationships with SoC vendors so that next-gen silicon arrives with your firmware already upstreamed and validated.