Edge Vision Engineer: Real-Time On-Device Inference

RAPSODO PTE. LTD.; Singapore; On-site; SGD 120,000 - 180,000; Full time

Job summary

Rapsodo PTE. LTD. in Singapore is hiring an experienced embedded software engineer to own on-device computer-vision inference pipelines.

You will implement and optimize end-to-end workflows from camera ingest to post-processing, targeting real-time performance on Linux-based edge devices. You will work closely with model and CV stakeholders to ensure edge-ready deployments, manage cross-compilation, and tune memory, latency, and power constraints for reliable field operation across our sports

Qualifications

• Master’s degree or PhD in Computer Vision, ML/DL, Electrical/CS or related field.
• Min 12 years of professional embedded software experience.
• Strong C++ (modern standards) with optional Rust; Linux expertise.
• Experience deploying CV/DL on edge hardware with quantization/optimizations.

Responsibilities

• Build and maintain production-grade embedded software for on-device CV inference (camera ingest, preprocessing, inference, postprocessing, telemetry) in C++ (Rust optional).
• Integrate edge runtimes and vendor SDKs for NPUs/MPUs.
• Profile and optimize latency, throughput, memory, power, and thermal use.
• Develop deployment-oriented model optimizations (quantization, operator fixes).
• Work on Linux-based embedded platforms: cross-compilation, packaging, and field deployment.
• Debug complex issues across kernel/user-space, I/O, and multi-threading.
• Collaborate with CV stakeholders to ensure edge-ready models and validation.
• Uphold engineering standards: code quality, CI, performance benchmarks, observability.

Skills

C++ (modern)

Rust (plus)

Linux fundamentals

Edge AI deployment

CV/DL inference concepts

Performance optimization

Education

Master’s degree or PhD in Computer Vision/ML/DL or related field

Tools

Perf

ftrace

eBPF

Yocto/Buildroot

CMake/Bazel

Job description

Rapsodo PTE. LTD. in Singapore is hiring an experienced embedded software engineer to own on-device computer-vision inference pipelines.

You will implement and optimize end-to-end workflows from camera ingest to post-processing, targeting real-time performance on Linux-based edge devices. You will work closely with model and CV stakeholders to ensure edge-ready deployments, manage cross-compilation, and tune memory, latency, and power constraints for reliable field operation across our sports