Condé Nast: AI-Powered Video Discovery and Content Intelligence on AWS

• 

  Condé Nast manages a large and growing library of premium video and digital content, creating challenges in efficiently discovering and reusing relevant assets.

• 

  Traditional discovery workflows made it difficult to locate relevant content using both semantic meaning and structured metadata.

• 

  Condé Nast needed to process a substantial historical content library while continuing to support new content as it was added.

• 

  Users needed search results that combined semantic relevance with metadata-based filtering and consistent scoring.

• 

  Failures during large-scale content ingestion needed to be detected, categorized, retried, and tracked.

• 

  The platform needed secure APIs and controlled access suitable for enterprise users and production workflows.

• 

  The initial search capability needed to provide a foundation for additional AI-powered content discovery and intelligence use cases.

• 

  Condé Nast wanted to extend AI capabilities beyond search into broader editorial and audience intelligence workflows.

NuVista AI designed and implemented an AWS-native content discovery foundation combining large-scale ingestion, semantic search, metadata filtering, secure APIs, and operational controls. The architecture was designed to support continued expansion into additional AI-powered content and intelligence capabilities.

Video Search & Content Ingestion

• 

  Built a scalable video search platform using Amazon S3 as the content source of truth and AWS services for ingestion, search, APIs, security, and monitoring

• 

  Implemented historical backfill and incremental ingestion workflows with idempotency, retries, reprocessing, and failure tracking

• 

  Combined semantic search with metadata filtering and normalized relevance scoring to improve content discovery and result consistency

Enterprise APIs & AI Expansion

• 

  Delivered secure versioned APIs with authentication, rate limiting, caching, error handling, monitoring, and infrastructure-as-code documentation

• 

  Extended the foundation toward GenAI image search, identity-based search, and AIHouse capabilities for editorial and audience intelligence

• 

  Validated a production-ready foundation for scalable AI-powered content discovery and created a path for expanding the platform into additional content intelligence workflows.

• 

  Delivered a secure, scalable platform for semantic and metadata-based discovery across Condé Nast's content library.

• 

  Processed ~150K historical records in approximately 4.5-5 days using eight concurrent requests.

• 

  ~150K records completed successfully, with remaining failures categorized and tracked for remediation.

• 

  Established monitoring, retry, reprocessing, failure tracking, secure access, and deployment documentation.

• 

  Created a foundation for additional GenAI search and content intelligence capabilities.

• 

  The initial search engagement established a broader foundation for continued AI and data modernization work.

• 

  Expanded the use of AWS-native services to support scalable AI-powered content and audience intelligence workflows.