Anthony Perez Email and Phone Number

Responsible for architecting, implementing, and maintaining economic well-being vertical products. This includes data analysis, data engineering, and data modeling / machine learning / computer vision on GIS data related to economic well-being to develop geo-referenced raster and vector products.• Designed and implemented the first iteration of the Atlas of Human Settlements (AHS) product, which detects buildings at the pixel level with Sentinel-2 satellite imagery using an encoder-decoder model written in pytorch. Specifically, AHS reports the fraction of each pixel’s area covered by buildings.• Lead the development of an internal MLOps platform for machine learning research and productization. Onboarded team members to the platform.• Lead the deployment of a workflow orchestration engine, Argo Workflows, on Kubernetes to manage data and machine learning pipelines with auto-scale of compute instances.• Containerized new and existing workflows with Docker for execution in Kubernetes (GKE).• Modernized practices across the team and mentored team members in adopting best practices. Practices include CI/CD with Google Cloud Build, Infrastructure as Code (IaC) with Terraform, containerization with Docker and Google Artifact Registry, role-based access with GCP IAM, and leveraging Spot/Preemptible instances in GCE.

One of the first three engineers hired at Atlas AI. Responsible for developing maps of socioeconomic well-being around the world. Developed Atlas AI's Asset Wealth Index (AWI) and Spending products. • MLOps - Designed and implemented configurable auto-scale of compute instances based on ML training and inference job requirements using Google Compute Engine.• MLOps - Storage, administration, and tracking of models, experiments, data, and code using BigQuery and Google Cloud Storage.• Machine Learning - Worked directly with clients to deliver solutions around socio-economic data.• Machine Learning - Developed Atlas AI's Asset Wealth Index (AWI) and Spending products from inception to release. Built computer vision models such as CNNs on GIS raster data using Tensorflow, Pandas, GeoPandas, NumPy, Scikit-learn, and GDAL.• Data Engineering - Built pipelines for model training and inference on GIS data using Spark (PySpark), Google Dataproc, BigQuery, and Google Earth Engine. The pipeline enabled the creation of core products by processing Landsat, Sentinel, and other public and proprietary image collections.• Data Engineering - Managed Atlas AI’s GIS data lake using Google Cloud Storage, BigQuery, PostGIS, GDAL, Rasterio, SQLAlchemy, and Google Earth Engine.

Research in Computational Sustainability. Specifically, Socio-Economic Modeling which involves mapping livelihoods in developing countries using large-scale machine learning models (graphical models and deep learning) and unconventional data sources. Presented a workshop paper in the NIPS 2017 conference on using publicly available Landsat 7 imagery to predict asset wealth in Sub-Saharan Africa with deep CNNs and transfer learning. Co-authored a nature communications article on the same subject. https://www.nature.com/articles/s41467-020-16185-w

Worked on designing assignment rubrics, enrollment, grading, and providing guidance during office hours for a project based course in which teams of 3-4 students tackled sustainable development problems with scarce or noisy data using deep learning techniques. Provided technical feedback both during presentations and in 1-on-1 settings. Prepared a VM disk on Google Cloud’s Compute Engine with CUDNN, TensorFlow, Numpy, etc. and provided support for students on creating VM instances and manipulating objects in Cloud Storage. Autumn and Winter 2017-2018 quarters.

Developed the CWC dashboard, which is an internal tool used by Citrix. The portion I was responsible for surfaced information relating to the Cloud Connectors used by Citrix Workspace Cloud. I worked as a full stack developer for this project, creating both a back-end API and front-end UI. The back-end was built with C# while the front-end used AngularJS, Twitter Bootstrap, Highcharts and plain HTML, JavaScript, and CSS.

Collaborated with a team that develops cloud solutions for Citrix Workspace Services. Worked on a cloud service that stored user subscriptions to workspaces as part of a cloud migration project which enables customers to migrate from their existing enterprise deployments to a cloud based deployment. Some responsibilities included:• Implemented designs, code reviews, debugged and integrated code for cloud services deployed in Microsoft Azure.• Wrote unit tests for various controllers.• Designed and implemented back end business logic for new features.• Utilized Agile development methodologies (SCRUM, Kanban, etc.)• Configured and deployed cloud services that communicated with other cloud services.• Created a cloud service that stored entities in an Azure table and allowed all CRUD operations (Create, Read, Update, Delete). The cloud service was a backend component, and part of the Citrix Workspace Services API.• Used Postman to manually test controllers, both deployed and locally with a debugger.• Used Azure Storage Explorer to view deployed table stores.• Instituted best practices and coding standards (code reviews, logging and instrumentation, unit tests, etc.)Gained experience in: .NET programming, C# and ASP, LINQ, Asynchronous programming (async/await), MVC web programming, Azure Table Store, IIS and web service hosting.