with the implementation of an end-to-end machine learning platform
“ Fluid has simplified out model building and deployment
process, and allowed for integration
across
our various platforms ”
Rainer Schilder - Lead Data Scientist
Case Study
Industry
Technology
Company Size
51 - 200 employees
Project
Length 7 months
Team Size
3.2 FTE
Problem Space
Data Science
AutoML
MLOps
Machine Learning
Technology Used
Python Kubernetes
FastAPI Kubeflow
KServe AWS
ScikitLearn Tensorflow
Business Context
At a fast-paced technology company, the product owners have a long list of backlogs driving high demand of
the data scientists attention. The complex technology landscape and disparate data systems create a high
friction environment for the data scientists to efficiently creating insights and models. The average time
from a problem to deployment takes 3+ months, resulting in the business losing competitive edge.
Challenge
Use cases take a long time to go from conception to production
Data scientists lack the bandwidth to support all
business use cases. This is compounded by the
complex development & deployment
environment, resulting in frustrated business &
technical team.
Solution
Creating an MLOps platform to support use case development & deployment
The platform allows the users to upload a dataset, create analysis and data transformation, run AutoML using
state-of-the-art machine learning
algorithms and deploy the model into production.
End-to-end MLOps Platform
The fast-iterative feedback loop improves efficiency and improves ROI
💰
20% cost reduction in operational
cost in deployed machine learning
🕑
Time saved for data scientist in
deploying models into production
🚀
Increased collaboration between
business users & technical team
