Python Data Science
Get there faster.
How one customer achieved 30x faster run-time completion with Wallaroo Labs.
“Achieving and exceeding the target run-time for a new data product we wanted to market was made possible through our partnership with Wallaroo Labs. Working with our data science team, Wallaroo Lab’s engineering team brought their expertise, a collaborative approach, and the Wallaroo framework to get the results we needed for our business!”
SVP Software Engineering
Customer Case Study
Customer challenge was to efficiently and effectively scale a new data application as a new product in the marketplace. The data scientists were running into issues with extremely long run-times, in excess of 3+ days. As the data volume and algorithmic complexity increased during their prototyping, things just got worse.
The application entailed taking device records data from customer provided CSV files and using Python data science libraries and tools to perform data linkages with another broader dataset of known devices.
The business opportunity to pursue larger more complex customer opportunities with larger data-sets was appealing, but the 3+ day run time was a non-starter for market viability.
The customer’s business and product leadership team had determined that they needed to be able to run the application, even in cases with significantly larger data-sets, in under 30 minutes to have a viable product for market.
Wallaroo Labs and the data scientist team went to work. Through a two-part process, the prototype run-time was reduced from 3+ days to under 12 minutes – meeting the customer’s goals for a viable new product offer for the marketplace!
The end result of a 12 minute run-time (far exceeding the minimal viable run-time of 30 minutes) was achieved through the Wallaroo Labs engineering team collaborative consulting engagement and the application of the Wallaroo framework’s parallelization and horizontal infrastructure scaling capabilities.
“Working with the Wallaroo Labs team and applying the Wallaroo framework was easy and fast via the Wallaroo Python API. We could use our existing algorithms as originally written in Python and easily take advantage of the framework’s horizontal scaling and resiliency.”
~ Data Scientist