Member of Technical Staff, New Grad

About us

We are building AI systems that can reason, use tools, and complete meaningful work in the real world. Our team works across model post-training, reinforcement-learning infrastructure, large-scale training, and product engineering. We believe the fastest path to more capable and reliable agents is an integrated loop: challenging environments, rigorous evaluations, efficient training, reliable inference, and products that make those capabilities useful.

You will join as a full member of the technical team—not as an observer. Depending on your strengths and the highest-priority problems, you may work on model post-training, RL environments, distributed systems, training or inference performance, backend services, APIs, SDKs, or developer tools.

This role is designed primarily for early-career engineers completing or recently completing a bachelor's or master's degree, or demonstrating equivalent ability through their work. We are looking for raw technical strength, evidence of initiative, and unusual learning speed. A strong academic record can be one signal, but school brand is not a requirement. We are equally interested in candidates whose ability shows up through open-source contributions, ambitious personal projects, research, internships, competitions, startup work, or products used by real people.

Candidates completing a PhD or bringing equivalent research depth may be a stronger fit for our Member of Technical Staff, Research — Early Career role.

• Own meaningful technical projects with clear user, research, or infrastructure impact; take them from problem definition through implementation, testing, rollout, and iteration.

• Write simple, high-quality code and learn the surrounding system deeply enough to make sound design and debugging decisions.

• Investigate ambiguous failures across unfamiliar layers, use evidence to narrow the problem, and ask for help in ways that accelerate the whole team.

• Contribute wherever the mission requires, which may include ML experiments, data or evaluation pipelines, distributed services, performance work, product features, and developer tooling.

• Build tests, metrics, documentation, and operational tools so that your work is reliable and maintainable rather than a one-time demo.

• Participate actively in design discussions and code reviews; communicate clearly, incorporate feedback quickly, and improve the quality of the team's decisions.

• Use modern AI development tools thoughtfully while verifying outputs and maintaining independent technical judgment.

• Share what you learn through internal notes, technical deep dives, examples, or open-source contributions where appropriate.

• Graduating soon or recently graduated with a bachelor's or master's degree in computer science, engineering, mathematics, or a related field—or equivalent ability demonstrated through your work. Roughly zero to two years of full-time industry experience is typical.

• Exceptional coding ability in at least one general-purpose language, supported by code, projects, internships, research, competitions, or other concrete evidence.

• Strong computer-science fundamentals, including data structures, algorithms, systems, and the ability to reason precisely about performance and correctness.

• A track record of learning difficult material quickly and completing ambitious work with limited structure.

• Grit, curiosity, humility, and a willingness to work hard on the highest-impact problem—even when it is outside your initial area of expertise.

• Clear communication and the ability to collaborate in a direct, fast-moving, low-ego environment.

• Meaningful contributions to a technically demanding open-source project or a substantial independent system used by others.

• Experience with ML systems, large language models, distributed systems, compilers, databases, networking, GPUs, or performance engineering.

• Strong results in research, competitive programming, systems or ML competitions, technical internships, or early-stage startup work.

• Evidence that you can turn an idea into a working artifact quickly, measure whether it works, and improve it from…