SDE, AI - DevOps, 10/2021 - 10/2022

During my time on the team, I was given the opportunity to take on a new workstream focused on measuring DevOpsGuru's efficacy. The idea was to match internal team tickets against produced insights in order to gain a clear picture of our performance. I accomplished this through the implementation of internal APIs for data collection. In addition, I delved into analyzing silencing anomalies and their impact on our efficacy. I created a framework and ran over 12,000 experiments to draw conclusions and improve our efficacy. Furthermore, I deprecated a component, resulting in a savings of at least $6000 per month.

After only a month of tenure, I was already participating in the on-call schedule. Throughout my time, I leveraged a range of technologies and languages, including AWS CDK, Lambda, DynamoDB, ElasticSearch, CloudFormation, S3, CloudWatch, Pandas, Python, Java, and Bash.

SDE, ML & Forecasting, 04/2020 - 07/2021

As one of the three SDEs in a research team, I had the opportunity to tackle several exciting projects. The first initiative involved benchmarking and empirical comparison of machine learning training jobs. Together with a colleague, I implemented and presented a Python CLI that allowed us to start benchmarking jobs using YAML configuration files. We leveraged this tool in combination with CodeBuild and SageMaker ProcessingJobs to create daily benchmarks of GluonTS algorithms.

In addition to this, I also worked on another Python CLI that made it easy for us to interact with SageMaker TrainingJobs. This tool allowed us to create summaries, replicate data structure for debugging, fetch logs, and more. Over time, I had the chance to assist the external and internal release cycle of GluonTS and contributed to a paper by preprocessing datasets.

Throughout these projects, I had the chance to work with a variety of technologies and languages, including Docker, Kubernetes, Boto3, SageMaker, S3, Click, Toolz, Python, and Bash.