Machine Learning Engineer, Orbital Sidekick

• If you are a smart-creative individual seeking to innovate and produce real, tangible value; take ownership of your role and position within a cutting edge tech company; and inform the future vision of an up-and-coming San Francisco start-up, you’ll fit right in
• 3+ years postgraduate experience as a MLEBS, MS, or PhD in Computer Science, Engineering, Physics, Mathematics or similar
• Python, Pytorch, Weights and Biases, ML Flow, DVC, Ray, Dask, data science stack
• Experience with deploying and monitoring models in clouds environments (AWS)
• Ability to work independently as well as collaboratively in a startup environment
• Comfortable with rapid decision-making and taking ownership of system design
• Excellent written and oral communication skills

Desirable

• We’re looking for a Machine Learning Engineer who will build out and design machine learning infrastructure for a team of data & remote sensing scientists, working collaboratively with the Analytics and Software teams to deliver data-driven products and solutions
• You will contribute to a cross-functional team that is responsible for the full algorithm development process, all the way from human annotation to integration with the platform
• Your role as Machine Learning Engineer means that you will be responsible for owning infrastructure for experiment tracking, model training/inference, dataset version control, and imagery annotation campaign creation and data extraction
• You will work collaboratively within the company to produce high-level engineering of both the data collection and information delivery platforms
• Setting up Machine Learning infrastructure including, model repositories, experiment tracking, workflow orchestration (Prefect, Airflow, Dagster)
• Optimizing resources for both development and production in the cloudDeploying models for both batch and online inference
• Establish model monitoring, model registries, experiment tracking, CI/CD, and continuous training
• Role will initially be focused on setting up MLOps and then grow into researching, developing, debugging, and deploying models to production using our hyperspectral data