Inference at scale
Latency budgets, batching strategies and what actually happened when teams moved models to the edge.
Technology and AI · Portland, Oregon
Where applied AI meets the people shipping it.
Two days of engineering-first talks on inference, evaluation and the unglamorous infrastructure behind useful models.
Latency budgets, batching strategies and what actually happened when teams moved models to the edge.
Practical eval harnesses, regression suites for prompts and how to argue with a leaderboard.
Pipelines, lineage and the retrieval layers that quietly decide whether a product works.
Why reliability beats novelty in production AI.
Benchmarks from three real deployments.
Bring a laptop. Leave with a CI job.
Guardrails, audits and rollback plans.
Lessons from rebuilding a vector stack twice.
Honest numbers on GPU spend.
Twelve speakers, five minutes each.
A candid retrospective from the program committee.
All speakers and organisations are fictional.
1120 NW Kestrel Ave, Portland, OR (fictional)
October 14 - 15, 2026
Registration · powered by Regform
This conference, its speakers, and its venue are fictional examples. The registration form below is a real, external Regform form. Information you submit may be saved, and any payment or email settings belong to that form. Please explore without submitting personal information or payments.
Open the same registration form on its own page.
Open hosted form ↗Opens in a new tab.
A real registration form, embedded right in this page.
View inline form ↓Open registration in an overlay without leaving the conference.
Form not loading, or prefer a separate page? Open the hosted form directly ↗