Discover the latest trends and best practices impacting data-intensive applications. Register for access to all 60+ sessions available on demand.
This session presents LinkedIn’s journey in transforming its recommendation systems from offline batch pipelines to real-time, AI-driven architectures. It highlights how frameworks like Entity-Based Recommender (EBR), Model Cloud, and Flyte-enabled experimentation unified retrieval and ranking layers. The talk focuses on improving freshness, latency, and efficiency through PyTorch-based online inference and GPU-optimized serving—delivering measurable impact on engagement, retention, and cost efficiency.
Nishant Lakshmikanth is an Engineering Leader at LinkedIn, with over 12 years of experience designing and leading large-scale distributed systems and infrastructure. He currently drives LinkedIn’s recommendation infrastructure, including the People You May Know (PYMK) system, a critical initiative responsible for generating millions in annual revenue, engaging over one billion members worldwide. His work integrates cutting-edge technologies like graph-based models, entity-based recommender systems, and advanced machine learning frameworks.
In the realm of machine learning infrastructure, Nishant has spearheaded the development of distributed training systems, real-time feature population pipelines, and solutions for remotely hosting complex models, enabling seamless integration of large-scale AI systems into production environments and ensuring their accessibility and reliability. He has also contributed to building highly reliable tracking systems, GPU optimizations, and cost-efficient large language model (LLM) deployments tailored for recommendation systems, advancing the scalability and efficiency of LinkedIn’s ML-powered products.
Before LinkedIn, Nishant held key engineering roles at Amazon Web Services and Cisco. At AWS, he contributed to Elastic Block Storage (EBS), where he designed a distributed volume placement system and optimized replication strategies, earning seven patents. At Cisco, he advanced video streaming and encoding technologies, demonstrating expertise in backend systems.
Nishant’s technical contributions extend beyond systems design to fostering innovation, mentoring engineering talent, and advancing engineering standards. His extensive experience with control plane for managing complex distributed systems, bootstrapping cloud services and building machine learning infrastructures places him at the forefront of innovation.
