Sign In

James Piette

Entrepreneur & Data Scientist in Finance & Sports

James Piette is currently the Managing Partner of Data Science at Blue Lacy Group. He has a robust background in data science and analytics, having previously held significant roles in various organizations. Notably, he served as a Partner and Director of Data Science at X10 Capital and was a Principal Data Scientist at LendUp before his tenure at Blue Lacy Group.1

Piette has also co-founded sports analytics startup Krossover Intelligence, showcasing his expertise in applying statistical methods to enhance athletic performance.2 His academic contributions include co-authoring research on the relationship between baserunning and batting abilities in baseball, published in the Journal of Quantitative Analysis in Sports.3

In addition to his professional endeavors, Piette's career reflects a blend of entrepreneurial spirit and technical proficiency, making him a notable figure in the fields of data science and analytics.

Highlights

Mar 20 · linkedin

Traditionally, I felt code deployment speed and rigor was a low priority for startups. I was very wrong.

This goes back to my origins as an ops engineer (read: statistician pretending to code). I would usually say to myself, "Self, the time taken to build the correct CI/CD pipelines will never outweigh the benefit that provides for a startup." Then, we would commit & deploy code to prod in lightning speed and pat myself on the back for time saved.

But slowly, as customers get added, the complexity compounds. Tiny changes you needed to make for one customer are missed for others. Versioning becomes difficult as backwards compatibility becomes a problem. Worse, the stress of managing it knowing something will break takes its toll on the team.

I'm obviously not saying anything that even the most junior developers don't already know. However, as business-minded executives, it's easy to forget that these small details matter.

Where I think execs and dilettantes like myself can make an impact is enforcing compromise. Maybe skip the unit tests and "60% code coverage." While blasphemous, the key in my mind is to focus on creating strong development environment. Then build simple tools to get that code deployed to production safely without much fuss.

Bugs will find a way in, but hey, you're working at a startup. You were up working anyway when the system went down at 9pm on a Friday night.

Mar 6 · linkedin

I doubt analysts will need AI to generate queries when IDEs like DataGrip already make it so easy.

Without an IDE, it is much harder for me to write queries. With one, the factors that limit my ability to produce proper data or analysis are

  • forming the right questions,
  • producing appropriate answers, and
  • selecting the applicable conditions.

None of those can be aided generative AI, at least from the few copilot-esque apps I've tried. But I am willing to be proven wrong!

Related Questions

What are James Piette's main responsibilities at Blue Lacy Group?
How did James Piette transition from Jindo Group to Blue Lacy Group?
What projects has James Piette led at Blue Lacy Group?
What is James Piette's educational background?
How has James Piette's career evolved in the data science field?
James Piette
James Piette, photo 1
James Piette, photo 2
Add to my network

Location

Greater Philadelphia