Machine Learning in Production: Challenges in combining Scala with Java
Submitted by Ravindra Guntur (@raviguntur) on Thursday, 5 February 2015
Startups attempting to deploy machine learning algorithms in production face many challenges. Most algorithms do not work as expected, and developers spend a lot of time and effort trying to integrate machine learning code with other parts of the system. This talk highlights some of the challenges that we faced and how we overcame them. We hope our experience will help other startups.
This talk will present our experiences when we deployed machine learning algorithms in production. We highight our experiences when we tried to combine Java implementations with Scala implementations in a SaaS Product. We describe why we choose the lambda architecture and how java and scala developers co-worked to build a complex system.
Ravindra is a lead engineer in the Knowledge Engineering group at Kaybus. He architects, designs and builds machine learning code for the cloud based SaaS product that Kaybus is developing. Ravindra has a PhD in computing from the Indian Institute of Science, and has been working in the software industry for over 10 years.