Supporting a Data Science Team
Beginning my journey as an MLOps Engineer
One of the hardest things about being a data scientist is that they expect you to also be a Software Engineer, a DevOps Engineer, a Data Engineer, and a Subject Matter Expert (if not more roles). I honestly enjoy doing all those things, when you have these expectations on you, you get to build a large breadth of experience. When working with large numbers of people that also have the same breadth it is a challenge to get models up in production and keep them stable in the long run. This year I decided I wanted to move into MLOps, a growing field adjacent to data science that applies DevOps to Machine Learning. I have now done so and joined as a Machine Learning DevOps Engineer (MLOps) at John Deere.
One of the reasons I picked to work at John Deere is the fact they were hiring people to work in MLOps. It is an incredibly positive signal to me that leadership understands that just throwing an additional data scientist at a challenge does not on its own solve problems. Data scientists should be supported properly and roles on the team are a reflection of that support. MLOps Engineers are there to help Data Scientists deploy and monitor models. Data Engineers are there to create feature stores that make data engineering tasks easier, faster, and more trustworthy. Data Catalysts (SME/Product Owners at…
