About the role
Placeholder: you’ll own parts of the codec stack that turns petabytes of gridded, imaging and sensor data into a fraction of the space, and the verification that proves the error bound holds.
What you’ll do
- Placeholder: train and benchmark neural codecs
- Design error-bounded and lossless modes
- Work with engineering to ship codecs to production
What you’ll bring
- Placeholder: 5+ years in applied ML, signal processing or compression
- Strong Python and PyTorch or JAX
- Experience with large scientific or imaging datasets
Nice to have
- Placeholder: published work on learned compression or SZ/ZFP-style compressors
Apply
Apply for this role.
Takes about five minutes. Have your CV ready.
Application sent.
Thanks for applying. We read every application and will be in touch either way.