Edge AI / Inference Engineer
What you will own.
Build the onboard inference pipeline within latency, power and thermal constraints. Profile and optimize the path from camera input to model output.
What we expect from you.
- Show measured optimization work on an embedded accelerator, including the methodology and any accuracy or stability cost.
- Debug problems across camera drivers, Linux services and inference runtimes rather than treating the model as an isolated component.
What you will do.
- Implement and instrument the camera-to-inference pipeline on the selected onboard compute platform.
- Profile memory movement, scheduling and runtime behavior under realistic sustained workloads.
- Evaluate model conversion and optimization changes against output correctness, latency, memory, power and temperature.
- Create repeatable device benchmarks, deployment packages and diagnostics that integration engineers can use.
What you bring.
- C++17/20, CUDA, TensorRT and ONNX Runtime.
- JetPack, DeepStream, GStreamer and V4L2.
- Nsight profiling, multithreading and high-throughput pipelines.
- Embedded Linux, Yocto or Ubuntu L4T.
Useful, not essential.
- Rust, Vulkan/OpenCL or hardware video encoding.
- VPI or image signal processing experience.
What good looks like.
- A reproducible device benchmark that identifies the actual bottleneck under sustained operation.
- An optimized pipeline with documented resource use, regression checks and a recovery path for runtime faults.
A conversation about your work.
Describe an edge pipeline you profiled. What was slow, how did you prove the cause and what changed after your optimization?
Use a non-confidential example. We care about your own contribution, the decisions you made and the evidence behind the outcome.
We are an early-stage, founder-led company. Expect hands-on work, changing priorities and direct responsibility for a clearly defined area. Write down decisions, raise risks early and distinguish demonstrated results from assumptions.
The experience range is a guide to scope and independence. Equivalent practical experience is welcome; required professional authorizations still apply. Hiring phases describe our plan, not 58 immediate openings. Location, employment terms and any role-specific access or field requirements are discussed before an offer.
Need an adjustment to the application or interview process? Contact contact@actprove.com.
Let’s talk.
Apply for Edge AI / Inference Engineer with relevant experience and examples of your work.
Please share only material you are permitted to disclose. Do not submit classified, controlled or employer-confidential information.