Multi-tenancy in Machine learning (the SaaS perspective)
Submitted by Swaminathan Padmanabhan (@swaminathankp) on Friday, 19 July 2019
Session type: Birds of a Feather session of 1 hour Session type: BOF session of 1 hour
Given the emergence of several SaaS product companies in India, there’s a lot of recent interest in provisioning ML capabilities over the cloud; and enabling SaaS customers make use of such capabilites through a self-serve model. These SaaS customers should be able to tailor the ML capabilities on-the-fly to suit their needs, e.g. they should be able to adjust confidence thresholds of a virtual chat bot as a fuction of their inbound query volumes and satisfaction levels. They should also be able to plug in custom datasets pertaining to their business and train the ML models further. We also strive towards improving bootstrap performance, i.e. a customer should be able to do minimal work and start consuming these ML features over the cloud. We also provision feedback loops with-in our products, and manage the real-time infrastructure for consuming feedback from customers (and customers of customers) at scale.
Freshworks is one of the largest SaaS startups in the country which offers cloud-based ominchannel customer engagement software. We are integrating AI/ML capabilites across our software suite which spawns across functions such as marketing, sales, customer success and support. We face several challenges on a day-to-day basis which concerns with enabling multi-tenancy in ML. We will also be discussing a few case studies based on our experience.
Multi-tenant ML models - enabling self-serve of ML capabilities at scale
Customizable ML models - enabling customers to seamlessly tailor models to fit their needs
Vertical-specific models - training models at an industry-level, making ML features available to new customers from day 0
Learning from continuous feedback at scale
Case studies - Automation through Virtual assistants (chatbots, intent detection, lead scoring)
Swaminathan Padmanabhan is currently Director, Datascience at Freshworks. He has been with Freshworks Inc since 2017 and has built a team of ~20 ML Engineers and Data scientists, who are based out of both the Bangalore and Chennai R&D centers. Prior to joining Freshworks, Swaminathan was heading the Datascience function at Olacabs, Bangalore; where he was leading a team of ~35 Research Engineers, Systems engineers and Data scientists. An alumnus of IIT Madras, he was also associated in the past with Inmobi and Yahoo Inc.
We’ll also have 1-2 external participants and 2-3 other participants from the Datascience team at Freshworks join the BoF session.
Other Freshworks participants:
Suvrat Hiran - Lead ML engineer (9632657694)
Varun Nathan - Lead Data scientist (9445356949)
External participants: TBD