Design and optimize large-scale pre-training systems for generative AI models, building distributed pipelines and improving training performance, reliability, and efficiency across hardware and software stacks.
Software Engineer, ML Ops
π Description
- Build and maintain pipelines for fleet field data, including rosbags, sensor logs, and telemetry.
- Create curated, versioned datasets and manage their storage, indexing, querying, and delivery.
- Set up training workflows and optimize cloud costs.
- Develop tools for fast data access, reproducible experiments, and automated evaluation.
- Generate metrics and diagnostics for dataset health, model performance, and pipeline reliability.
π οΈ Requirements
- Bachelor's or Master's degree in Computer Science, Robotics, Data Engineering, or a related field
- Strong Python proficiency and working knowledge of ROS2
- Working knowledge of docker and other DevOps tools
- Familiarity with cloud storage and compute (AWS - S3, EC2, etc.)
- Understanding of ML workflows and dataset versioning
- Master's in Computer Science, Robotics, or a related discipline
- 2+ years of MLOps or data infrastructure experience, ideally in robotics or autonomous systems
- Experience with Weights & Biases, rosbag data, and large-scale sensor datasets
- Working knowledge of C/C++
- Experience supporting perception or ML research teams
Full job description
Who We Are
AeroVect is transforming ground handling with autonomy, redefining how airlines and ground service providers around the globe run day-to-day operations. We are a Series A company backed by top-tier venture capital investors in aviation and autonomous driving. Our customers include some of the worldβs largest airlines and ground handling providers. For more information, visit www.aerovect.com.
You will
Build and maintain data pipelines to ingest field data (rosbags, sensor logs, telemetry) from our fleet
Convert raw field data into curated, versioned datasets for the perception team and own dataset management - storage, indexing, querying, and vending datasets
Set up training workflows and optimize cloud costs
Build tooling to accelerate perception engineers' workflows - fast data access, reproducible experiments, automated evaluation pipelines
Generate metrics and diagnostics to track dataset health, model performance, and pipeline reliability
You have
Bachelor's or Master's degree in Computer Science, Robotics, Data Engineering, or a related field
Strong Python proficiency and working knowledge of ROS2
Working knowledge of docker and other DevOps tools
Familiarity with cloud storage and compute (AWS - S3, EC2, etc.)
Understanding of ML workflows and dataset versioning
We Prefer
Master's in Computer Science, Robotics, or a related discipline
2+ years of MLOps or data infrastructure experience, ideally in robotics or autonomous systems
Experience with Weights & Biases, rosbag data, and large-scale sensor datasets
Working knowledge of C/C++
Experience supporting perception or ML research teams
Please note this role will be based onsite in Toronto
Similar jobs
Search more Machine Learning Engineer jobsMachine Learning Engineer 1
Own the end-to-end machine learning lifecycle for misdirected email detection, from data and feature engineering through deployment and monitoring, improving detection efficacy through experiments and production-ready systems.
Machine Learning Engineer II (ROA + LTV ML)
Develop and scale production machine learning models for credit decisioning, including default prediction and approval decisions, while building training, monitoring, and data systems with engineering, credit, and product teams.
Build and evolve AI/ML infrastructure that enables product teams to develop, deploy, and scale AI-powered applications, supporting model training and serving, data workflows, orchestration, and evaluation.
Build and evolve AI/ML infrastructure that helps product teams develop, deploy, and scale AI-powered applications, supporting model workflows and customer-facing experiences across Thumbtack.
