Jordan O Brown

Architected agentic orchestration and RAG retrieval pipelines in Python to increase task completion accuracy by 27.5% and answer relevance by 24.3%, while reducing latency and inference cost through EKS autoscaling and Terraform-driven deployments.

Designed RAG and retrieval architectures and built production LLM applications with LangChain and custom orchestration, improving client conversion and response rates and implementing evals frameworks to reduce prompt defects.

Built embeddings-based retrieval and RAG prototypes with FAISS and designed data/indexing pipelines to improve search relevance and freshness, while implementing observability and inference orchestration to cut costs and MTTR.

Built retrieval and indexing pipelines for IP search using embeddings and FAISS, implemented grounding and hallucination mitigation, and deployed ML services on Kubernetes to improve search speed and reduce incidents.

Developed ETL pipelines and data quality checks and built search indexing with Elasticsearch to improve dataset freshness and recall, while implementing evaluation metrics that improved ranking performance on validation sets.

Completed a Master of Science degree at the University of Amsterdam in 2017 focusing on advanced topics in the field.