Praveen Mallya
Employers on the program
- 2026DDN1 talk
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
- STAC Summit
Talks
| Date | Event | Conference | Role | Title | Links |
|---|---|---|---|---|---|
| May 20, 2026 | STAC Summit New York, Spring 2026 | STAC Summit | speaker | Predicting AI Inference Efficiency Under Realistic Mixed Workloads as DDN Standard AI inference benchmarks separate cache hits from misses into sequential phases, force uniform request sizes, and provide no explicit control over cache…Show lessStandard AI inference benchmarks separate cache hits from misses into sequential phases, force uniform request sizes, and provide no explicit control over cache hit rate — limitations that prevent them from modelling production traffic where heterogeneous document lengths and mixed cache states arrive interleaved. We show how to generate realistic interleaved request streams with configurable hit rate probability, heterogeneous document lengths with weighted distributions, and controlled concurrency. |
Frequent co-speakers
No one else spoke at two or more of the same events.
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.