Dan Feigin
Employers on the program
- 2026NVIDIA1 talk
Events1
Conferences1
Talks1
Years active2026
The year around now
Talks in the past 12 months and announced talks ahead, one row per conference, January to December. Hover a mark for the talk.
Appearances by year
- AI Engineer (World's Fair, Summit and regional editions)
Talks
| Date | Event | Conference | Role | Title | Links |
|---|---|---|---|---|---|
| Sep 23, 2026 | AI Engineer Paris 2026 | AI Engineer (World's Fair, Summit and regional editions) | workshop | From Minutes to Seconds: Killing the Cold Start Tax on LLM Inference as NVIDIA Spinning up an LLM replica can be painfully slow. Before a fresh inference server instance outputs a single token, it has to move weights onto the GPU, spin up…Show lessSpinning up an LLM replica can be painfully slow. Before a fresh inference server instance outputs a single token, it has to move weights onto the GPU, spin up CUDA, sit through compilation, CUDA graph capture, engine initialization, and more. That's tens of minutes of idle GPU time, but that's the cheap part. |
Frequent co-speakers
| Speaker | Shared events | Shared talks | Last together |
|---|---|---|---|
| Ron Kahn | 1 | 1 | 2026 |
Speakers with the same name are merged into one entry, so this page may combine more than one person. Affiliations are shown as each conference listed them.