Vela (YC W26) in San Francisco is seeking candidates who are intellectually curious and experienced in building AI agents. Responsibilities include solving unique problems, creating agents for daily decisions, and understanding customers’ needs to develop practical solutions. Ideal candidates have a strong initiative, test-driven development experience, and an interest in machine learning and reinforcement learning. Join a vibrant team that values creativity and collaboration while developing cutting-edge technology.
Qualifications
Ultra intellectually curious; participates in competitions or does fun research.
Strong agency and initiative; tackles issues personally.
Collaborative; enjoys working with friends and humor.
Responsibilities
Solve complex problems without known solutions.
Build decision-making agents used daily.
Deliver usable products to customers swiftly.
Engage with customers to understand and meet their needs.
Oversee systems from research to production.
Skills
Curiosity
Problem-solving
Test-driven development
Machine Learning knowledge
Reinforcement Learning interest
Job description
Who are you?
Ultra intellectually curious (you are active in competitions/championships such as ICPC, IOI, Putnam, etc.; or do research for fun; or win hackathons)
Have crazy amounts of agency (passion projects for life; fix things that annoy you)
Love working with friends (extremely strong opinions, weakly held; have a great sense of humor)
Build agents and are an AI power user (but can think for yourself)
Fast learner
What we would love for you to become (or you already are)
Curious about Machine Learning and Reinforcement Learning
Agent evaluator
Test-driven developer
What do you get to do?
Solve problems that don't have a known solution yet.
Build agents that make thousands of context-dependent decisions daily.
Ship things that real paying customers use the next day.
Talk to customers, understand their needs, and build.
Own entire systems end-to-end: from the research that informs an approach, to the architecture, to the code, to watching it work in production.
Work at the frontier of what agents can actually do reliably, not what demos look like on Twitter.