By Ben Sykes In a previous post, we described how Netflix uses Apache Druid to ingest millions of events per second and query trillions of rows, providing the real-time insights needed to ensure a high-quality experience for our members. Since that post, our scale has grown considerably. With our database holding over 10 trillion rows and …
Enterprises often struggle to onboard new team members at scale. Human resources (HR) teams spend time on manual tasks that delay productivity, such as processing documents to answering repeated questions about benefits and policies. For organizations with many new hires, these steps make it harder to keep onboarding consistent and compliant. Organizations lose substantial amounts …
Read more “Build AI-powered employee onboarding agents with Amazon Quick”
Synchronizing the Senses: Powering Multimodal Intelligence for Video Search By: Meenakshi Jindal and Munya Marazanye Today’s filmmakers capture more footage than ever to maximize their creative options, often generating hundreds, if not thousands, of hours of raw material per season or franchise. Extracting the vital moments needed to craft compelling storylines from this sheer volume of …
Read more “Powering Multimodal Intelligence for Video Search”
In today’s agentic AI environments, the network has a new set of responsibilities. In a traditional application stack, the network mainly moves requests between services. But as discussed in a recent white paper, Cloud Infrastructure in the Agent-Native Era, in an agentic system the network sits in the middle of model calls, tool invocations, agent-to-agent …
Read more “Envoy: A future-ready foundation for agentic AI networking”
Despite their sophisticated general-purpose capabilities, Large Language Models (LLMs) often fail to align with diverse individual preferences because standard post-training methods, like Reinforcement Learning with Human Feedback (RLHF), optimize for a single, global objective. While Group Relative Policy Optimization (GRPO) is a widely adopted on-policy reinforcement learning framework, its group-based normalization implicitly assumes that all …
Read more “Personalized Group Relative Policy Optimization for Heterogenous Preference Alignment”
By Renata Teixeira, Zhi Li, Reenal Mahajan, and Wei Wei On January 26, 2026, we flipped an important switch for Live at Netflix: all Live events are now encoded using VBR (Variable Bitrate) instead of CBR (Constant Bitrate). It sounds like a small configuration change, but it required us to revisit some of the foundational assumptions …
Read more “Smarter Live Streaming at Scale: Rolling Out VBR for All Netflix Live Events”
Evaluating single-turn agent interactions follows a pattern that most teams understand well. You provide an input, collect the output, and judge the result. Frameworks like Strands Evaluation SDK make this process systematic through evaluators that assess helpfulness, faithfulness, and tool usage. In a previous blog post, we covered how to build comprehensive evaluation suites for …
Read more “Simulate realistic users to evaluate multi-turn AI agents in Strands Evals”
Building the perfect bra takes thousands of data points. That’s why Honeylove isn’t just another intimates brand. We’re a technology company that happens to make exceptional bras, tops, shapewear, and bodysuits. Technology shapes everything we do, from how we iterate garments based on customer feedback to how we optimize sizing across those thousands of data …
Read more “How Honeylove boosts product quality and service efficiency with BigQuery”
Monitoring competitor prices is essential for ecommerce teams to maintain a market edge. However, many teams remain trapped in manual tracking, wasting hours daily checking individual websites. This inefficient approach delays decision-making, raises operational costs, and risks human errors that result in missed revenue and lost opportunities. Amazon Nova Act is an open-source browser automation …
Read more “Automating competitive price intelligence with Amazon Nova Act”
As AI workloads transition from experimental prototypes to production-grade services, the infrastructure supporting them faces a growing utilization gap. Enterprises today typically face a binary choice: build for high-concurrency, low-latency real-time requests, or optimize for high-throughput, “async” processing. In Kubernetes environments, these requirements are traditionally handled by separate, siloed GPU and TPU accelerator clusters. Real-time …
Read more “Run real-time and async inference on the same infrastructure with GKE Inference Gateway”