Tesseract: ML in Container Infrastructure
A hybrid elastic cloud and on-prem infrastructure is the new normal. ML has been called upon to enable overlooked efficiency gains leveraging opportunities in this new dynamic. Tesseract at PayPal is our ML enabler for our infrastructure to find optimality in resource efficiency and cost with strong reliability and resiliency guarantees. In this talk, we will introduce one such ML enabler from Tesseract driving pool right sizing at all times to meet anticipated demand. We present a Hybrid Deep Learning and Statistical approach to model anticipated demand. An accurate measure of incoming demand enables us to be right-sized while keeping guarantees on reliability, resiliency, and availability. Our experience dealing with demand volatility and how we temper them to enable actionability with tradeoffs will be the centerpiece of this talk.
Aashish Sheshadri is a Staff Machine Learning Engineer at PayPal. Where he enables ML for the Infrastructure. Applying ML to infrastructure efficiencies, security hardening and accelerating MLOps. Before PayPal, Aashish spent 4 years in Robotics and NLP research at UT Austin and CMU.