.webp?width=250&height=78&name=DC-Logo%20Type_%20Invert(1).webp "Data Council")
Alyssa Frazee tells us about the unicorn data skills she's honed on the job.
One thing that Alyssa Frazee loves about her work at Stripe is that, like someone with traditional data science skills, she gets to build machine learning models. "Oh, the rapture," cries Alyssa the data scientist!
But there's another part of Alyssa's job that she relishes just as much - she's also the hands-on engineer that ships the models she builds as production software systems. "Real artists ship," says Alyssa the (Jobsian) data engineer.
Often in data teams, these two roles are perceived as dichotomies - the work done by the 'data scientist' in many companies is clearly separate from the work done by the 'data engineer.' But Stripe isn't your typical company, and Alyssa isn't your typical engineer.
But what are there advantages to this approach? And is a hybrid 'machine learning engineer' job function a model that more companies should consider?
Meet Alyssa Frazee
Alyssa is a machine learning engineer at Stripe, where she builds models to prevent online credit card fraud. Prior to joining Stripe, she completed a PhD in biostatistics before falling in love with programming at the Recurse Center.
When the real world is different than your textbooks
One of the times when Alyssa appreciates the special power of being a double-threat is when a real-world project throws her a curveball; she runs up against a problem that, on the surface, appears theoretically simple but in the end turns out to be deceptively complicated to answer. This is where having a practical approach to problem solving - plus a special toolkit she's developed (ability to add error bars to complex metrics, a way to monitor models in production when you can't observe the outcome, and a method to explain decisions made by black-box models) - serves her better than a purely theoretical approach.
In her talk at DataEngConf, Alyssa will take us on our tour of what it's like to live in her hybrid world between data science and engineering, and she'll show us how to use the tools from her bag of tricks along the way. Using a plethora of examples from her payment fraud detection work at Stripe, she'll uncover practical ways that engineers can both implement and troubleshoot machine learning models, and give us an appreciation for the practical power behind the unique skills of 'machine learning engineer.'
(1).jpg?width=1920&name=Pano2(1)(1).jpg "Data Council Blog Signup")