Scholar
Avner May
Google Scholar ID: Gx5baHUAAAAJ
Staff Research Scientist at Together.ai
Citations & Impact
Citations
575
H-index
13
i10-index
15
Publications
20
Co-authors
14
list available
Contact
Publications
Resume (English only)
Research Experience
- Staff Research Scientist at together.ai.
- Research Scientist at Google's speech recognition group (2020–2023).
- Postdoctoral scholar in Prof. Chris Ré's group at Stanford University (2018–2020).
- Software Development Engineer at Microsoft for two years prior to PhD (Seattle, WA).
- Conducted social network analysis research (2011–2013) under Prof. Augustin Chaintreau.
Background
- Research interests center around designing simpler, better understood, and more efficient machine learning models.
- During PhD, demonstrated that kernel approximation methods can match fully-connected deep neural networks on nonlinear speech recognition tasks.
- Recently focused on understanding what makes approximate feature representations perform well on downstream tasks, in both kernel approximation and word embedding compression.
- This understanding aids in efficiently selecting or designing feature approximations and navigating trade-offs between computation, memory, and performance.
- Prior to ML, conducted two years of research in social network analysis, studying whether platforms like Facebook or Twitter efficiently deliver content of interest to users.
