Software Engineering Daily Production-Grade AI Systems with Fred Roma
17 snips
Jan 27, 2026 Fred Roma, SVP of Product and Engineering at MongoDB, a veteran in cloud and data management. He talks about the complex AI stack: LLMs, embeddings, vector search, caching, and observability. He covers schema evolution in the LLM era, Voyage AI’s multimodal embeddings and rerankers, and how data platforms must adapt for production-grade AI systems.
AI Snips
Chapters
Transcript
Episode notes
Prioritize Simplicity, Accuracy, Evolvability
- Simplify your AI data stack and prioritize accuracy and evolvability as primary goals.
- Design for cost-effectiveness and the ability to swap models or tools as the ecosystem changes rapidly.
Schemas Lose Durability In LLM Era
- Schemas are far less durable in the LLM era because models and integrations change fast.
- Roma highlights that MongoDB's document model and JSON make schema evolution easier for AI workloads.
Co-Locate Search With Operational Data
- Combine operational data, search, vector search and model optimizations within the same platform to reduce integration complexity.
- Avoid stitching many separate systems; instead, bring search and vector search near your operational data.
