Neeraja graduated with a PhD in Computer Science from University of California, Berkeley. Her thesis was on automatic resource management in the datacenter and the cloud. She is now a post-doctoral researcher in the Computer Science Department at Stanford University where she continues to work on distributed systems, cloud computing, and machine learning. Neeraja received her masters in Computer Science from the Indian Institute of Science, Bangalore, India.
Most of Neeraja’s research straddles the boundaries of systems, and Machine Learning (ML). Advances in Systems, Machine Learning (ML), and hardware architectures are about to launch a new era in which we can use the entire cloud as a computer. New ML techniques are being developed for solving complex resource management problems in systems. Similarly, systems research is getting influenced by properties of emerging ML algorithms, and evolving hardware architectures. Bridging these complementary fields, her research focuses on using and developing ML techniques for systems, and building systems for ML.
“Traditional resource management techniques that rely on simple heuristics often fail to achieve predictable performance in contemporary complex systems that span physical servers, virtual servers, private and/or public clouds. My dissertation work brings the benefits of data-driven models to resource management of such complex systems. I argue that the advancements in ML can be leveraged to manage and optimize today’s systems by deriving actionable insights from the performance and utilization data these systems generate. To realize this vision of model-based resource management, I dealt with the key challenges data-driven models raise: uncertainty in predictions, cost of training, and generalizability from benchmark datasets to real-world systems datasets.
In future, I plan to continue working on problems that further the over-arching vision of an easy-to-use and cost efficient cloud. In theory, such a cloud would offer management-less and fine-grained consumption-based access to satisfy users’ cost and performance goals. As a postdoc, I am building a simple interface for serverless computing systems that hides the complexity of resource allocation from users while satisfying their cost-performance goals. Further, in the context of ML workloads, I envision a model-less interface for serving inference from ML models; such an interface allows users to specify high-level goals in terms of accuracy, latency, and performance SLOs. Working with students at Stanford, I built a model-less serving system to support this interface with novel model-selection and autoscaling mechanisms. I believe, taking these steps will enable a fully automated cloud, meeting the requirements of existing and emerging new workloads.”