Energy-Efficient AI: Building Sustainable Data Pipelines for the Future

Artificial Intelligence has become one of the largest consumers of computational resources, with training and inference workloads growing exponentially. While AI fuels innovation across industries, its energy footprint raises concerns around cost, scalability, and sustainability. This talk will explore practical strategies for building energy-efficient AI systems, focusing on Green Data Engineering practices that optimize resource usage without compromising performance. Topics will include: • Dynamic batching and KV caching for reducing inference overhead • Sparse neural networks and structured pruning for lightweight models • Carbon-aware scheduling to align compute with renewable energy availability • Federated learning and edge deployments for reducing data transfer energy costs • Sustainability maturity models for organizations adopting Green AI practices The session blends academic research, industry case studies, and real-world implementations from large-scale data ecosystems to highlight how AI practitioners can design pipelines that are not only fast and scalable but also environmentally responsible. Attendees will walk away with actionable frameworks to measure, monitor, and minimize the carbon footprint of AI workloads—making sustainability a core design principle, not an afterthought.
Who is this presentation for?
Researchers and Industry Professionals
Prerequisite knowledge:
AI
What you'll learn?

Bhanu Prakash Reddy Rella is a Senior Data Engineer at Meta, where he designs and optimizes large-scale data pipelines powering AI-driven marketing insights and decision-making. Previously, he worked as a Lead Data Engineer at Walmart, building cloud-native data infrastructure and performance pipelines, which gave him extensive experience in big data ecosystems, distributed processing (Spark, Hive, BigQuery), and cloud platforms. He is also the Founder of The Green AI Initiative, a global community advancing sustainable AI practices. Bhanu is pursuing a Doctor of Business Administration (DBA) at Golden Gate University, San Francisco, with a thesis on Energy-Efficient AI and Green Data Engineering. His contributions span patented innovations in sparse neural networks, IEEE/Elsevier/Scopus-indexed research publications, and he authored the book Energy-Efficient Computing for Modern AI Applications With over a decade of experience across AI, ML, and large-scale data systems, Bhanu brings a unique perspective at the intersection of innovation and sustainability, inspiring the AI community to align cutting-edge advancements with planetary well-being.