David Vieira

I’m a Senior Applied AI Engineer with 6+ years of experience building large-scale travel search, ranking, and AI systems across Perplexity AI, Expedia Group, and Airbnb. I focus on end-to-end intelligent systems that integrate retrieval, reasoning, and real-world actions—so answers aren’t just accurate, they’re actionable in production.

At Perplexity, I owned development of the travel vertical, building a RAG-driven, agentic LLM system with hybrid BM25 + vector search and multi-model routing for reliable tool/API execution. I engineered production LLM inference services (PyTorch, with critical components migrated to Rust) on AWS GPU clusters via Kubernetes, achieving ~30% lower median latency and driving measurable improvements in answer accuracy (~15%), task reliability (~20%), uptime (>99%), and latency (sub-500ms P95). Previously at Expedia and Airbnb, I scaled hybrid search/RAG pipelines, ranking systems, and agent orchestration using Elasticsearch, Kafka, and AWS, repeatedly improving relevance, task completion, and conversion through rigorous evaluation and A/B testing.