Nick Graeff Email & Phone Number

Las Vegas, Nevada, United States

• Planned, documented, wrote, and tested dozens of AB tests across the site for features such as autocorrect, swatches, improved result ordering, etc.
• Migrated various services from manual deployments to full CI/CD in AWS
• Designed and built entire pipelines and tier 1 services across the site
• Identified bottlenecks in latency across the site and worked with other teams to resolve
• Optimized build speeds across multiple teams, assisting with getting local builds working and parity with production systems
• Contributed to Amazon internal packages used across the org

Las Vegas, Nevada Area

• Designed and distributed Android library (LaunchKey Authenticator SDK) via Github/Maven that enables clients to use multi-factor authentication without any of the complications of setting it up themselves
• Automated end-to-end testing on several Android apps by building a testing suite using Appium, Docker, and Cucumber BDD tests, reducing QA time several magnitudes
• Kickstarted CI/CD of apps using Jenkins pipelines and AppCenter, including running merge checks (static analyzers and unit tests) for PR’s and integration tests prior to deployment, increasing customer confidence in.
• Modularized several portions of the SDK for reuse in projects shared with other teams (i.e. encryption, networking, utils)
• Updated several Android apps regularly to use the latest LaunchKey Authenticator SDK's features among other frameworks used in the apps (Dagger, Firebase, JUnit, Mockito, AndroidX, etc.)
• Added new features to the LaunchKey Authenticator SDK such as authorization by Country/State/Zip location, conditional authorization based on the aforementioned feature, authorization by detection of connected.

Las Vegas, Nevada Area

• Lead project to completion focused on Spectral Clustering, an unsupervised machine learning algorithm used to separate loosely connected objects from each other, resulting in a 10% improvement in the quality of region.
• Implemented parallel image processing algorithms (morphology, connected components) in C++ and CUDA, typically improving performance/speed of those functions by 30%
• Optimized CMake build configuration to cut build times from 6 minutes to 30 seconds
• Refactored code, removed duplicate/similar functionality in many places, eliminated memory leaks, and rewrote certain processes to run in parallel
• Upgraded team’s build environment from VS2010 to VS2017 and other tools and libraries such as Qt4 to Qt5, Eigen to MKL, CUDA 8 to CUDA 9, and Python 3.3 to Python 3.6 enabling faster development turnaround
• Automated data collection tasks and built development tools in Python to measure performance and analyze trends in data which allowed for incremental improvements to classification