Technical Talks

ABOUT THE TALK

• GenAI Applications

When Hunch's viral LinkedIn year-in-review AI generator reached 300,000 users processing 1+ trillion tokens in two weeks, their multi-model architecture faced extreme scaling challenges. This case study reveals how a simple prototype evolved into a production-scale AI system overnight. Discover Hunch's technical blueprint featuring multiple LLM orchestration across OpenAI, Anthropic, and Google models, critical infrastructure scaling solutions, and how they achieved 85% cost reduction through optimized model selection and prompt engineering. Learn from their 26 rapid iterations that simultaneously improved output quality while decreasing costs. This presentation shares practical patterns for AI workflow orchestration balancing quality, cost, and reliability at scale. Gain actionable engineering strategies for building resilient, scalable AI applications that maintain performance under unpredictable growth, plus vital lessons about system failure points when success arrives faster than expected.