Pavithra Gunasekaran is an applied scientist on the Azure Data Centers team at Microsoft. She has over seven years of experience in large-scale data analysis, predictive modeling, and system-performance analytics across product, healthcare, and e-commerce. At Microsoft, she built the constrained optimization engine for space, power, cooling, and network placement across about 550 data centers, which informed $4B / 800 MW of global over-allocation potential. She also built production ML forecasting models (risk quantile, drift, and trajectory) for compute and power utilization under failure scenarios. She architected automated ML pipelines that process more than 78 billion rows of hourly telemetry, with CI/CD and auto-retraining, which cut model refresh time by 32%. She leads a team of three and has mentored eight engineers.
Before Microsoft, she was a senior data specialist/scientist at BlinkRx. There she built production forecasting models for operational workload and anomaly-detection pipelines for distributed claim-processing systems, and she automated onboarding and reporting pipelines for pharma clients, including Pfizer and Sanofi. At Groupon, she delivered a recommender system for more than 60 million active users. At the Indian Space Research Organization, she analyzed high-frequency satellite telemetry to detect anomalies and predict system failures, and she delivered a dashboard that was patented with ISRO. She has also worked at IDG Communications, Accenture, and Deloitte.
Pavithra holds a Master’s in Business Analytics and Data Science from the University of Texas at Dallas and a Bachelor’s in Computer Science and Engineering from Dayananda Sagar College of Engineering. She co-authored an IEEE research paper on vehicle detection and received the Microsoft Lab Most Innovative App Award.