Senior Perception Engineer, Edge ML for Smart Home

Sauron Industries, Inc.; San Francisco (CA); On-site; USD 220,000 - 260,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

Sauron Industries, Inc. seeks a Senior Perception Engineer to own the end-to-end perception stack for home security sensors. You will shape sensing, inference, and deployment strategies, combining vision, ML, and systems work across embedded devices and cloud components.

You’ll drive cross‑functional collaboration, set performance targets, and ensure the perception system improves in accuracy, latency, and reliability for real-world environments.

Qualifications

• 5+ years building production systems for hardware products.
• Experience with perception or ML in safety-critical domains.
• Strong C++20, memory and concurrency discipline.
• Hands-on with vision, tracking, and multi-camera setups.
• Experience with PyTorch, TensorRT, and edge inference.
• Fluent Python and strong debugging/observability skills.

Responsibilities

• Own perception architecture, performance, latency, and cost tradeoffs.
• Define sensing requirements across product generations with hardware.
• Set the boundary between systems and models in the stack.
• Lead evaluation, regression harness, and model selection criteria.
• Fuse sensor observations robust to occlusion and noisy data.
• Ensure trustworthy deployment with failure detection and observability.
• Drive dataset creation and evaluation tooling for continuous improvement.

Skills

C++20

Perception/ML for hardware

GStreamer

Computer vision

PyTorch

Python

CUDA

Edge inference

Multi-camera tracking

Tools

Abseil

gRPC

CMake

TensorRT

DeepStream

NVIDIA Jetson

Job description

Sauron Industries, Inc. seeks a Senior Perception Engineer to own the end-to-end perception stack for home security sensors. You will shape sensing, inference, and deployment strategies, combining vision, ML, and systems work across embedded devices and cloud components.

You’ll drive cross‑functional collaboration, set performance targets, and ensure the perception system improves in accuracy, latency, and reliability for real-world environments.