Blog

••7 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

••7 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Kiril Urbonas·2

••7 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Kiril Urbonas·5

••8 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

••8 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Kiril Urbonas·2

••8 months ago

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Kiril Urbonas·3

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Architecture Review: LLM Gateway Design for Multi-Provider Inference

We started routing 90% of LLM traffic through a small internal gateway. The gateway wasn't planned — it emerged from solving the same problem in 5 places. Here's the shape it took.

Kiril Urbonas·10

Real-World RAG Incidents: Lessons from a Production Rollout

A field report from rolling out retrieval-augmented generation in production, including cache bugs, bad embeddings, and how we fixed them.

Kiril Urbonas·6

Embedding Models Comparison: Choosing the Right Model for Your Use Case

We benchmarked six embedding models on the same retrieval task. The results that surprised us, and how we'd pick today.

Kiril Urbonas·22