Research Fellow

Abundant•San Francisco, CA

About The Position

Abundant is building the NVIDIA of training data, focusing on the critical need to scale data for AI model training. The company's founding team has extensive experience from leading tech companies like Waymo, Google, Mercor, and AWS, with a proven track record in deploying deep learning models, training state-of-the-art models for self-driving cars, and scaling data pipelines. Abundant addresses the growing scarcity of training data, which is essential for AI advancements beyond current capabilities, aiming to enable AGI, ASI, and physical intelligence by providing the necessary data for domain expertise and multimodal AI. The company collaborates with top AI labs, startups, and enterprises.

Requirements

Research expertise in a related field (e.g., CS, ML, NLP) ideally with that work published at a major conference (NeurIPS, ICML, ICLR).
Velocity to master complex fields and a proven track record in the productization of research, encompassing high-stakes evaluation, large-scale benchmarking, and product feature development.
Proven experience shipping research directly to production and live systems, specifically focusing on advanced post-training, distillation, and high-stakes evaluation methodologies.
Experience with large-scale agent systems: orchestration frameworks, tool APIs, distributed execution, observability, and logging infrastructure.
A thoughtful, strategic perspective on the societal and safety impacts of deploying general-purpose AI systems.
Experience advising on or shaping governmental policy related to AI safety and governance (e.g., House of Lords or Online Safety Bill initiatives) is required.

Responsibilities

Architect and execute a core research agenda to discover simple, generalizable ideas that advance model reasoning and intelligence at scale.
Own the full research-to-production lifecycle and ensure rapid deployment of your work in live systems.
Partner with the world's most advanced AI research teams to design, engineer, and iterate on high-impact datasets and large-scale benchmarking efforts that shape how frontier models behave, specifically focusing on critical alignment, safety, and defining optimal reward signals.
Autonomously identify, scope, and manage long-running research projects, choosing the most impactful problems that are critical to scaling data for AGI/ASI.
Collaborate closely with engineering teams on data pipelines, internal tooling, and high-performance deep learning algorithm implementations.