Machine Learning Engineer

About the role

Mach Industries develops autonomous defense platforms for contested environments where traditional sensing is unreliable. You will manage the end-to-end data and training backbone required to power vision and sensor models across our product lines. This role requires a generalist approach to build the infrastructure that enables rapid iteration and real-time model deployment on embedded hardware.

Key facts

What you'll do

• Develop and maintain data ingestion, curation, and labeling workflows for flight, simulation, and hardware-in-the-loop data.
• Manage distributed multi-GPU training infrastructure, experiment tracking, and model registries.
• Optimize and deploy models for real-time edge inference on Jetson-class hardware using quantization and pruning techniques.
• Build and refine models for detection, segmentation, tracking, and multi-sensor fusion.
• Create synthetic data pipelines to address long-tail scenarios and bridge the gap between simulation and reality.
• Implement runtime monitoring for drift detection and performance metrics to inform retraining cycles.
• Collaborate with autonomy, embedded, and flight-test teams to transition models from prototypes to deployed systems.

Requirements

• Proficiency in Python for machine learning and production-grade C++ on Linux.
• Experience building full-cycle ML pipelines including dataset versioning, augmentation, and reproducible training.
• Hands-on experience fine-tuning CNN and Transformer architectures using PyTorch.
• Background in deploying models to embedded GPU hardware with strict latency and power constraints.
• Familiarity with MLOps tools, SQL, Parquet, and CI-based validation.
• BS, MS, or PhD in CS, EE, Robotics, or equivalent experience with a history of shipping production models.

Nice to have

• Experience with synthetic data generation using Unreal or Isaac and domain randomization.
• Knowledge of EO/IR imagery and multi-modal sensor fusion.
• Familiarity with CUDA, ROS 2, and NVIDIA Jetson deployment pipelines.
• Experience with active learning, data mining, and distributed training frameworks.
• Proficiency in Docker and Rust.

Skills & tools

• Python, C++, PyTorch, SQL, Parquet, TensorRT, ONNX Runtime, CUDA, ROS 2, Docker, Rust.

Practical notes

This position involves work subject to U.S. export control laws, including EAR and ITAR. Employment offers may depend on your eligibility to access controlled technology without an export license. Mach uses E-Verify to confirm work authorization. The compensation package includes base salary, equity, health insurance, and retirement plans.