Research Scientist Intern
Pluralis Research · San Francisco; Australia
Extracted from the posting
6-month fixed-term research internship for current PhD candidates, focused on publishing tier-1 ML papers in Protocol Learning. Work covers communication-efficient training, fault tolerance under node churn, heterogeneous compute, and robustness to malicious participants. Team is remote with main teams in Australia and North America, with access to significant compute and senior mentorship.
Pluralis Research works on Protocol Learning: training and serving large models in a fully decentralized way on small consumer-grade devices connected via the internet. Despite being dismissed as infeasible, we have made significant advances on this problem, most recently Agora, a permissionless run that pretrained an 8B model from scratch on consumer GPUs spread over the internet, with no single participant ever holding the full weights (tech report). While many of the core research problems have been solved, Protocol Learning unlocks a series of new challenges. For the mission in full, read A Third Path: Protocol Learning.
This setting breaks nearly every assumption of datacenter training and inference: communication-efficient training across different parallelism axes, fault tolerance as nodes join and drop mid-run, heterogeneous compute and networks, and robustness to malicious participants. Our published methods include Subspace Networks, Factored Gossip DiLoCo, AsyncMesh, and Sentinel.
As a Research Scientist Intern you join us for a 6-month, fixed-term research internship during your PhD, focused on publishing. You work on the problems that stay open as we push from the 8B run toward frontier scale, with access to significant compute and focused mentorship from senior scientists. These are foundational papers up for grabs.
Key Responsibilities
- Publish in Tier-1 venues: Conduct novel research in Protocol Learning with the explicit goal of publishing in tier-1 ML conferences (NeurIPS, ICML, ICLR).
- Own a real problem: Pick a question that blocks Protocol Learning at scale and answer it — the internship is scoped so a foundational paper is a realistic outcome, not a stretch goal.
What We're Looking For
- Publication track record (required): Current PhD candidate with at least one publication in top-tier ML venues (NeurIPS, ICML, ICLR).
- Research focus: You work in a core technical area relevant to frontier models.
- Technical depth: Strong theoretical understanding of deep learning and distributed systems principles.
- Implementation skills: Proficiency in PyTorch and experience with large-scale training infrastructure.
- Mission alignment: You believe Protocol Learning is the viable third path for collective, trustless, and sovereign AI.
FYI's
- This is a 6-month, fixed-term research internship. We only hire in Australia and the US, and visa sponsorship is limited to these countries.
- We work remotely across the world, with the main teams in Australia and North America. You'll need to be comfortable working across timezones.
- Applicants must have professional-level English proficiency (written and spoken).
- Recruiters: we aren't looking for agency support at this time. We'll reach out if we need help.
We are backed by Union Square Ventures and other tier-1 investors, and we are a world-class, deeply technical team of ML researchers. Pluralis is unapologetically ideological. We believe AI, and the world, end up on a better path if we succeed in implementing the protocol for intelligence. If this resonates, please apply.
