Real-Time Edge ML Engineer for Autonomous Vision

Delos; San Francisco (CA); On-site; USD 120,000 - 220,000; Full time

Job summary

Mach Industries is seeking a Machine Learning Engineer to own data and training infrastructure for its AI-forward autonomy stack. You will build scalable pipelines, manage synthetic data, and enable real-time edge inference on embedded hardware in flight.

You will work across perception, localization, embedded and flight-test teams to ship production-grade models. The role emphasizes end-to-end ML pipelines, distributed training, and hands-on deployment on embedded GPUs, balancing latency,

Qualifications

• Strong generalist software engineering with Python ML and production C++ on Linux.
• Experience building ML data and training pipelines end to end.
• Hands-on PyTorch training and fine-tuning for detection/segmentation/tracking.
• Edge deployment on Jetson-class hardware with TensorRT/ONNX.
• Data/MLOps infrastructure including dataset/versioning, CI, and scalable multi-GPU training.

Responsibilities

• Own and evolve data and training infrastructure for the autonomy team.
• Scale training/eval infrastructure, experiment tracking, and model registry.
• Deploy and optimize real-time edge inference on Jetson-class hardware.
• Improve models across portfolio for detection, tracking, and ATR.
• Generate and manage synthetic data at scale using simulation.
• Instrument runtime health, drift detection, and retraining loops.
• Collaborate with perception, localization, embedded, and flight-test teams.

Skills

Python for ML and tooling

C++ on Linux

Profiling & optimization

ML data pipelines

PyTorch training & fine-tuning

Edge deployment (Jetson)

TensorRT/ONNX Runtime

MLOps (SQL/Parquet, CI)

Education

BS/MS/PhD in CS/EE/Robotics or equivalent

Tools

NVIDIA Jetson

TensorRT

ONNX Runtime

CUDA

Docker

ROS 2

Parquet/SQL

Job description

Mach Industries is seeking a Machine Learning Engineer to own data and training infrastructure for its AI-forward autonomy stack. You will build scalable pipelines, manage synthetic data, and enable real-time edge inference on embedded hardware in flight.

You will work across perception, localization, embedded and flight-test teams to ship production-grade models. The role emphasizes end-to-end ML pipelines, distributed training, and hands-on deployment on embedded GPUs, balancing latency,