Virgin Hyperloop – Software Engineer
-
After MSc graduation, decided to stay in the US and continue my work at Hyperloop. I worked for Machine Intelligence Team, Data Automation & Insights group, focused on generative transportation system design. I was wearing many hats: Software Architect, Back-End, Data Engineer.
- designed and co-developed a cloud simulation run tool in Go (run & monitor) and Python (analyze), using EC2, MLflow and Databricks. Allowed for massively scaled simulation & batch results (e.g. running and analyzing thousands of parallel, hours-long simulations on EC2/S3).
- developed advanced simulation metrics in python and dashboards, allowing C-level executives to summarize simulation runs and get actionable insights about Hyperloop system design for a given location.
- sped up python analytics from hours to minutes with cloud Spark and Koalas (we wrote it up with Databricks),
- trained & deployed 3 demand prediction models using keras & Horovod, ARIMA, Prophet. Researched the effect of weather, events, holidays on transportation demand. Improved Hyperloop capacity utilization by 70% (in simulation).
- presented my team’s work at conferences, blog-posts & webinars, reaching > 50k technical viewers.