Research Scientist Intern
Pluralis Research · Remote · San Francisco · Australia ·
COMPENSATION
€•• – ••k
Listed on the employer's job board (Ashby)
- 6-month fixed term
- Fully remote
- PhD candidates only
- Publish in NeurIPS/ICML
- Visa sponsorship (US/AU)
Join Pluralis Research to tackle challenges in decentralized AI, specifically training large models on consumer-grade devices. This fixed-term internship offers the chance to solve open problems in fault tolerance and communication efficiency while aiming for foundational publications.
What you'll do
- Conduct novel research in Protocol Learning with a focus on publishing
- Target tier-1 ML conferences such as NeurIPS, ICML, or ICLR
- Identify and solve specific bottlenecks blocking scale in decentralized training
- Produce foundational papers as a primary outcome of the internship
What you'll need
- Current PhD candidate with at least one top-tier ML publication
- Research background in core technical areas relevant to frontier models
- Strong theoretical understanding of deep learning and distributed systems
- Proficiency in PyTorch and experience with large-scale training infrastructure
- Professional-level written and spoken English proficiency
- Alignment with the mission of trustless and sovereign AI development
Stack
- PyTorch
How you'll work
- Fully remote work across global time zones
- Main teams located in Australia and North America
- Comfort with asynchronous collaboration required
- Hiring restricted to candidates in Australia and the US
What you get
- Access to significant compute resources
- Mentorship from senior scientists
- Backed by Union Square Ventures and other tier-1 investors