This position requires that the candidate selected be a US Citizen and currently possess and maintain an active Top Secret security clearance.

Join a sizeable team of data scientists, research scientists, and machine learning engineers that develop computer vision models on overhead imagery for a high-impact government customer. We own the entire machine learning development life cycle, developing models on customer data:

- Exploring the data and brainstorming and prioritizing ideas for model development
- Implementing new features in our sizable code base
- Training models in support of experimental or performance goals
- T&E-ing, packaging, and delivering models

We perform this work on both unclassified and classified networks, with portions of our team working on each network.

We seek a new team member to work on the classified networks. Three to four days a week, you would travel to the customer site in Northern Virginia to perform tasking as described below. Weekdays when you do not travel to the customer site, you would work from your local Amazon office. You would work collaboratively with teammates to use and contribute to a well-maintained code base that the team has developed over the last several years, almost entirely in python. You would have great opportunities to learn from team members and technical leads, while also having opportunities for ownership of important project workflows. You would work with Jupyter Notebooks, the Linux command line, Apache AirFlow, GitLab, and Visual Studio Code. We are a very collaborative team, and regularly teach and learn from each other, so, if you are familiar with some of these technologies, but unfamiliar with others, we encourage you to apply - especially if you are someone who likes to learn. We are always learning on the job ourselves.


Key job responsibilities
With support from technical leads, carry out tasking across the entire machine learning development lifecycle to develop computer vision models on overhead imagery:

- Run data conversion pipelines to transform customer data into the structure needed by models for training
- Perform EDA on the customer data
- Train deep neural network models on overhead imagery
- Develop and implement hyper-parameter optimization strategies
- Test and Evaluate models and analyze results
- Package and deliver models to the customer
- Incorporate model R&D from low-side researchers
- Implement new features to the model development code base
- Collaborate with the rest of the team on long term strategy and short-medium term implementation.
- Contribute to presentations to the customer regarding the team’s work.