We believe that data science should be treated as software engineering.
It’s easy and fun to ship a prototype, whether that’s in software or data science. What’s much, much harder is making it resilient, reliable, scalable, fast, and secure. We’ve spent five years building and running our platform, and want to share some thoughts about what we’ve learnt along the way. Above all else, we believe that data science should be treated as software engineering.
Our mission at Ravelin is to generate accurate predictions of risk at speed and scale, and we apply our predictions to the task of stopping fraud online. We are purposely not dogmatic about the methods employed to make those predictions, and use a combination of Expert Rules, Graph Databases and Machine Learning (ML). …
This article attempts to offer some of the rules for a productive, cohesive and enjoyable working environment for tech teams. It is a semi-working article that I will add to over time.
Tidy, readable, simple code is paramount. It makes reviewing PR easier, it makes coming back to the code after 18 months easier, it makes your peers like you more.
Constructive conversations over background criticism. Call it out when you see it (even if in other parts of the company). If you feel the need to complain, make a difference and give feedback instead.
During busy times, when you are deep in code and stress is mounting, it is easy to forget to Treat Others as You Would Expect to be Treated and be “approachable”. …