Research Scientist, Robotics

About Basis

Basis is a nonprofit applied AIresearch organization with two mutually reinforcing goals.

The first is to
understand and build intelligence. This means to establish the mathematical principles of what it means to reason, to learn, to make decisions, to understand, and to explain; and to construct software that implements these principles.

The second is to
advance society’s ability to solve intractable problems
. This means expanding the scale, complexity, and breadth of problems that we can solve today, and even more importantly, accelerating our ability to solve problems in the future.

To achieve these goals, we’re building both a new technological foundation that draws inspiration from how humans reason, and a new kind of collaborative organization that puts human values first.

Research scientists lead Basis’ efforts to develop a deeper understanding of the conceptual, mathematical, and computational principles of intelligence.

We are looking for people who are technically excellent, and who value probing concepts at their foundations. Our research scientists/engineers aspire to do rigorous, high-quality, robust science, but are not afraid to tinker, make mistakes, and explore radically different ideas in order to get there.

Basis is a collaborative effort, both internally and with our external partners; we are looking for people who enjoy working with others on problems larger than ones they can tackle alone.

We are searching for robotics expertise across multiple projects, both active and in development.

Active projects:

Science advances by discovering useful abstractions. MARA operationalizes this insight into software agents. Instead of passively absorbing data, MARA agents propose hypotheses, run physical or simulated experiments, and revise models of the world until a compact, coherent theory emerges. Early results show MARA systems solving previously unresolved (ARC) Abstraction and Reasoning Corpus tasks; widely considered among the hardest AI benchmarks today.

We are opening a new branch of the project to bring MARA into the physical world, with embodied agents that learn from experience, build internal representations, and apply "everyday science" human-like reasoning to solve problems.

Read the latest from MARA on our blog.

R-ADA aims to automate robotics, co-designing robot morphology (hardware) with controllers (software). We are treating the robot design as a sequential decision process, akin to human designers. A multimodal reasoning system designs robots as CAD programs, deploying both tools (e.g. FEM analysis) and internal world models, to understand what changes it should make and why. Candidate robots are evaluated in simulation using cross-morphology meta-policies, efficiently estimating their optimal performance across various tasks.

A probabilistic back-end updates its understanding of the physics gap with every real-world test. The payoff is a pipeline to quickly design optimal robots tailored to specific tasks, on the fly.

• Have a PhD (or equivalent research experience) in a relevant area such as Robotics, Computer Science, Electrical Engineering, or a related quantitative field.
• Have hands‑on experience with real‑world robots, including:
  • Working with robotic hardware, APIs, and real‑time control systems.
  • Integrating and utilizing camera systems and other sensors for robotic perception.
  • Developing and deploying robotic control policies for manipulation tasks.
• Have experience with relevant aspects of machine learning for robotics, particularly some of the following:
  • Deep learning architectures for vision and language processing in robotics.
  • Imitation learning and reinforcement learning for robotic control.
  • Techniques for robust generalization to unseen tasks, objects, and environments.
• Have demonstrated an ability to conduct first‑class scientific research, published in venues such as CoRL, ICRA, RSS, NeurIPS, ICML, or in technical reports.
• Have demonstrated an ability to write excellent software.
• Be excited about solving real‑world…