Zbigniew Mastylo

The CTO hired me to help them scale and replace a poorly implemented, crash-prone Ruby-based system with a robust, fault-tolerant, and scalable data pipeline.
I didn't get many requirements, except that I entirely owned the product and made decisions on tech stack, design, and implementation. I took on the challenge, conducted quick research, and found an excellent Open-Source library, Luigi, from Spotify.
I read the docs, looked at APIs, and wrote a sample pipeline deciding the library would be a good run for us. I presented my findings to the team and instructed them to keep working on their current tasks while I set up the project.

I wrapped Luigi in simple code, hiding a few complexities of the framework while exposing base interfaces to implement new behavior. There was no need to know or understand what Luigi was.

The new pipeline implementation was relatively simple: figure out what tables to join to get data and how to process that data. On average, implementing a new data pipeline took three to eight days. I led and managed the project from inception to product delivery, helping Bivarus scale and acquire at ten times valuation. This was an excellent run for my team and me.

I led the design and development efforts of the food ordering back end. The system was developed in Python using FastAPI, Postgres, Celery, and many Python libraries. It was fully integrated with Deliverect for real-time updates of restaurant menus.

I designed and implemented an extensible pathway testing framework at healthcare startup Privis Health (former Acupera). The framework allowed nurses and medical staff to create and execute tests against long-term care plans created using Acupera Pathway Language.