Building Production RAG Pipelines That Actually Work
Most RAG demos look impressive. Most RAG production systems disappoint. The gap between them comes down to retrieval quality, not model capability.
- AI Engineering
- LangChain
- RAG
- Azure OpenAI
Blog
Thoughts on data architecture, AI engineering, and building modern data systems at scale.
Why Your Data Lakehouse Needs a Streaming-First Design
Batch pipelines still dominate most lakehouses, but the cost of that choice compounds over time. Here's the architectural case for building stream-first from day one.
- Data Architecture
- Apache Kafka
- Streaming
- Azure