Resume

First joining Righthand Robotics, I executed performance and behavior tests on the product, similar to a quality engineer. From the behavior tests, I understood how our product interacts on a system level, if a feature is behaving the way its expected to, are there any bugs with a feature, and how to properly document and asses on a system level, what constitutes a behavior test case. From performance testing I learned what is important when assessing performance of the product (e.g. cycle time, different modes of failure, etc.) so that the team can better understand what is noise (or outside of our control) versus what was affected by this release. Then, for the majority of my time there, I was part of the Vision & Planning scrum team, which dealt with CV and ML related problems. I provided testing guidance, developed testing plans (which for the initial release of a feature were robust and well documented) and any behavior tests needed for future release testing in order to catch any regressions (making sure the feature is integrated well within the system). I was a part of the design meetings, team meetings, and pull requests for many CV, ML, and product features. I provided insight of robot configuration, robot environment, troubleshooting/debugging errors, pull request reviews, and testing guidelines. I created and further developed python scripts, unit tests, and repository files that were needed to release a new feature or test it. When possible, I utilized python and ansible to automate certain tests and behaviors, such as verifying robot information, health, configuration, output from nodes or messages, etc. to better understand what is going on. This required a deeper understanding of our system (on a software level) of what is going on in the state machine, how different messages, classes, and planners interact with each other, and how can we extract information from them to verify certain behaviors. I had to read through and understand different parts of the repository in order to successfully extract info that would be needed for a test, and could possibly be used for other testing. I also executed other tasks, such as uploading production images into our ML training dataset, develop tests that output metrics around the image and depth quality of our cameras, participate and present in R&D meetings around recent advances in AI, and prototype new feature ideas as part of hack-a-thons, whether it's simulating our robots, designing a new ML pipeline to feed and use data, or possibly using vision transformers for segmentation.