Git as the CMS: How This Blog Publishes Itself
Why I publish this blog by committing markdown to a git repository instead of running a blog server, and how the pipeline keeps the site up even when everything else is down.
Search architecture and relevance engineering, from concept to measurable results.
I am Jason Stubblefield, a search and discovery leader with more than twenty years across the full search stack. I design and ship production search platforms on OpenSearch, Elasticsearch, and Solr, with deep hands-on experience in vector retrieval, RAG pipelines, relevance experimentation, and LLM-assisted ranking. This site expands on my professional work: the research platforms I build, the open source tools I maintain, and my writing on search architecture.
From SolrCloud to OpenSearch: A Practitioner's Migration Guide
Real-world insights and practical steps for migrating from Apache SolrCloud to OpenSearch, covering challenges, solutions, and lessons learned from actual migration experience.
Why Search Will Still Be Relevant in the AI Era
Examining the continued importance of search engines and information retrieval systems in the age of large language models and AI assistants.
Summarization as Compression in Search Engine Architecture
Exploring how text summarization techniques can be used as a form of compression to improve search engine performance and relevance.
Featured Project
LLM connected recipe discovery and management platform.
Areas of Expertise
- • Search Architecture: OpenSearch, Elasticsearch, and Solr Cloud at scale
- • Vector Retrieval: KNN, embeddings, and visual search
- • RAG Pipelines and LLM-Assisted Relevance Tuning
- • Relevance Experimentation and A/B Testing
- • Cloud Infrastructure: AWS, GCP, Terraform, and Ansible
- • Engineering Leadership and Team Building