Large clusters for small models
Conference Context
- Date/time: 2026-07-01 · 1:55pm-2:15pm
- Track/room: Inference · Track 9
- Speaker(s): Daniel Svonava
- Session type/status: session · confirmed
- Track: Inference
- Room: Track 9
- Session type: session
- Status: confirmed
Session Description
Small task-specific models are cheaper, faster and narrowly better than the frontier. But a wide catalog of small models is tricky to serve - dedicated worker pools sit idle, top-down request routers choke up on the huge volume of small requests, your users bring 100s of LoRAs.. In this talk we show how we serve 1M tokens per second with small models, how we architect our cluster for maximum throughput AND minimum latency and how we apply autoresearch to rewrite our inference code to support 10+ new models a week.
Media Evidence
No related AI Engineer channel video found yet.
Evidence Graph
This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.
Media Signals
No linked video, transcript, or slide source has been attached yet.
Agent Reading Notes
Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.
Transcript Status
No official session recording transcript was found by exact title match on the AI Engineer YouTube channel during this run.
People
Notes
- Pending transcript synthesis when an official recording or confirmed matching video is available.
Synthesis
Synthesized Breakdown
Large clusters for small models ## Conference Context - Date/time: 2026-07-01 · 1:55pm-2:15pm - Track/room: Inference · Track 9 - Speaker(s): Daniel Svonava - Session type/status: session · confirmed - Track: Inference - Room: Track 9 - Session type: session - Status: confirmed ## Session Description Small task-specific models are cheaper, faster and narrowly better than the frontier. But a wide catalog of small models is tricky to serve - dedicated worker pools sit idle, top-down request routers choke up on the huge volume of small requests, your users bring 100s of LoRAs.. In this talk we show how we serve 1M tokens per second with small models, how we architect our cluster for maximum throughput AND minimum latency and how we apply autoresearch to rewrite our inference code to support 10+ new models a week. ## Media Evidence No related AI Engineer channel video found yet.
Speaker And Company Context
- Daniel Svonava — CEO and co-founder at Superlinked.
Topics Covered
Derived Links And Source Material
Novel Concepts / Clever Methods
- No highlighted novel concept has been detected yet.
Evidence Boundary
This synthesis is based on the official schedule and linked source pages. It should be revisited when exact session recordings or transcript-backed secondary sources are available.