Sign InOpen Brain
AI EngineerVideoSource Linked

Building Turbopuffer: Gergely Orosz (@pragmaticengineer ) × Simon Eskildsen (CEO)

Turbopuffer’s early architecture shows a useful agent-era discipline: ship the smallest system that tests demand, then optimize against measured hardware constraints. The transcript offers principles, not benchmarks.

AI Engineer · Aug 3, 2026
Open Source Open MarkdownOpen JSON
Source Summary

Turbopuffer’s first vector-search design put clustered vectors and centroids in files, fetched nearby clusters from S3, and used an Nginx cache on **one server**. Its founders drew on **eight years at Shopify**, and Cursor later adopted the database after other Postgres-based approaches.

Practical Implication

Builders working with agents should keep generated architecture proportional to the current problem: start with inspectable primitives, measure latency and cost, and add bespoke machinery only after demand appears. Hardware flexibility also matters when CPU capacity is constrained.

Agent-Ready Context
Turbopuffer’s first vector-search design put clustered vectors and centroids in files, fetched nearby clusters from S3, and used an Nginx cache on **one server**. Its founders drew on **eight years at Shopify**, and Cursor later adopted the database after other Postgres-based approaches.

Builders working with agents should keep generated architecture proportional to the current problem: start with inspectable primitives, measure latency and cost, and add bespoke machinery only after demand appears. Hardware flexibility also matters when CPU capacity is constrained.

The supplied transcript is incomplete and conversational. It does not provide comparative performance, cost figures, migration details, or enough evidence to generalize Turbopuffer’s architecture to other workloads.
Connected Context · Feed7 Judgment

This grounds the candidates’ “smallest useful system” principle at the infrastructure layer: Turbopuffer began with files, object storage, caching, and one server rather than premature distributed machinery. It adds latency, cost, and hardware availability as concrete gates for architectural escalation, while the incomplete evidence prevents treating that path as a reusable vector-search blueprint.

Context Map
infracodingdata#harness-engineering
Uncertainty
The supplied transcript is incomplete and conversational. It does not provide comparative performance, cost figures, migration details, or enough evidence to generalize Turbopuffer’s architecture to other workloads.