AI-Powered Search at Scale

At Frame.io (Adobe), we built AI-powered search to help creative professionals find the exact shot or moment across millions of assets. This required two AI integrations: an LLM for natural language queries and vector embeddings for semantic visual search. Our LLM component uses Meta’s Llama 4 via AWS Bedrock, translating queries like “videos of people on beaches from last week” into OpenSearch DSL. For visual search, we generate vector embeddings using Adobe’s SearchCut API (CLIP-based models), enabling similarity searches without metadata. The challenge? Video embedding takes 30+ seconds per asset. Our Oban infrastructure—already split across four instances due to database contention—couldn’t handle this load without impacting critical jobs. Our solution: a multi-stage Broadway pipeline consuming from SQS. A router pipeline handles authorization and versioning, fanning out to dedicated embedding and transcription workers. This provides independent scaling, traffic buffering, and complete isolation from our job system. You’ll learn: practical LLM integration patterns, when to choose Broadway over Oban, and how to build observable AI pipelines in Elixir.

Key Takeaways:

• Practical AI integration patterns they can apply immediately
• Decision framework for Oban vs Broadway
• Real production architecture from a major Adobe product

Target Audience:

• Primary Audience
• 1. Backend Engineers building search or AI features
• Teams integrating LLMs (OpenAI, Bedrock, Anthropic) into Elixir applications
• Developers implementing semantic/vector search capabilities
• Engineers working with OpenSearch, Elasticsearch, or similar
• 1. Engineers managing high-volume async workloads
• Teams hitting scaling limits with Oban or database-backed job queues
• Developers evaluating Broadway for production use cases
• Anyone processing external API calls at scale
• Secondary Audience
• 1. Tech leads and architects
• Making build-vs-buy decisions for AI/ML integrations
• Evaluating infrastructure patterns for async processing
• Planning feature-gated rollouts of AI capabilities
• 1. Oban users curious about alternatives
• Understanding when Oban is (and isn’t) the right tool
• Learning complementary patterns that work alongside Oban