Bolt is one of the fastest growing companies in Europe, and our engineering team is orders of magnitude smaller than those of rivals. Which means that to be competitive, we need to be efficient with our time and resources. Algorithms, including machine learning, are among the tools we use to improve the efficiency of our platform and provide the best experience for our clients. To do so, data scientists need great tools to develop and give life to new ideas.
That’s why we’re building data science platform, that does a lot of heavy lifting in areas of data collection and indexing, preprocessing (aka building features) and automating machine learning models lifecycle – training, deployment, and monitoring.
In March 2019 I gave at talk at NorthStar AI meetup in Tallinn about how engineering supports data science and enables faster growth at Bolt. This was also the first time we’ve talked publicly about the architecture of our feature store, and how it evolved to the default way of adding new features for ML models and fraud checks.
You can check out video below. Enjoy and feel free to comment and share your experience in building tools for data science and machine learning!