{"schemaVersion":"jobsearcher.job.v1","id":"d75761188c0fbe916d7cdcee","url":"https://jobsearcher.com/jobs/d75761188c0fbe916d7cdcee","canonicalUrl":"https://jobsearcher.com/jobs/d75761188c0fbe916d7cdcee","title":"Founding Engineer","description":"Kiteis where robots get trained. Augmentation, training, evaluation, and deployment in one platform. Whoever owns this layer decides how the next decade of robots gets trained. We intend to be that team.Robots are about to be deployed in labs, factories, warehouses, and homes. The models are getting very smart. The tooling is not keeping up. Teams write their own pipelines, run evaluations manually, and lose most of their time to infrastructure instead of robot behavior.We are 2 founders. You would be the third person and the first founding engineer. The culture we are building: robot arms, 3D printers running, prototypes in various states of working. Watching a robot do the thing you meant it to do never gets old. Some days that happens here, some days at a customer site. Everyone helps and there is no ego in the room.The roleYou will decide how the platform is architected. Decide it, build it, and own it in production. This is hands on from day 1.Work you will pick up in the first 6 months:Training infrastructure: GPU orchestration, scheduling, resource allocation, memory management, cost per jobThe execution layer that runs augmentation, training, and evaluation jobs across cloud providersSimulator integrations and the dataset pipelineThe platform API and the surfaces on top of it: web, CLI, MCPTalking to the researchers and engineers using all of the above and fixing what they tell you is brokenWhat we are looking forAttributesWe filter on 3 things.Learning & execution speed.Robotics changes every week. A new dataset format, a new policy architecture, a new simulator. We need someone who reads the paper on Monday and has a prototype by Tuesday.Communication.Topics in robotics can get very abstract quickly, you will be in front of customers, on calls with their engineers, and you have to make hard ideas simple in real time.No ego. Weare willing to learn what we do not know yet to make what seems impossible possible. We are honest with each other, take feedback, and are optimists.Ideal ExperienceYou have architected a system from scratchGPU orchestration, scheduling, resourcing, memory. You know why a training job is slow and you know what it costsRobotics software, or infrastructure for robotics teamsYou have trained robots multiple times and are familiar with the spaceComfortable across the stack. Python, PyTorch, Docker, and enough front end to ship a feature end to endReasons not to take this jobIt is hard. Most of what we are building has no established answer in robotics yet. You will be working things out from first principles and some of them will be wrong.The field moves fast enough that work you did 3 months ago may need to be thrown away.Your role will be diverse. Infrastructure one week, a customer integration the next, front end after that.There is no schedule. There are customers, deadlines we set ourselves, and things that break.SF is our central office and we value in-person time. If you are based abroad we will arrange something personally, but you will still travel often.You will travel to customer sites. Robots are physical and you cannot debug a deployment over Slack.You will be client facing. You will sit 1 to 1 with engineers from other teams and work on their problems.What you getMeaningful equity. You are the first engineer and the number reflects that.Competitive cash compensation.Full ownership of the platform architecture, with no legacy system and no one to ask for permission.The work itself. Robotics training infrastructure is being defined right now and very few people are working on it. What you build here changes how the next generation of robotics teams trains autonomous robots.","company":"Kite Ml","rawCompany":"kite ml","city":"Sonoma","state":"CA","isRemote":false,"isActive":false,"createdAt":"2026-09-03T13:00:22.745Z","occupations":[{"code":"17-2199.08","title":"Robotics Engineers","slug":"robotics-engineers"},{"code":"17-3024.01","title":"Robotics Technicians","slug":"robotics-technicians"},{"code":"15-1299.08","title":"Computer Systems Engineers/Architects","slug":"computer-systems-engineers-architects"}],"industries":[{"code":"541512","title":"Computer Systems Design Services","slug":"computer-systems-design-services"},{"code":"541511","title":"Custom Computer Programming Services","slug":"custom-computer-programming-services"},{"code":"333248","title":"All Other Industrial Machinery Manufacturing","slug":"all-other-industrial-machinery-manufacturing"}],"jobPosting":{"@context":"https://schema.org","@type":"JobPosting","title":"Founding Engineer","description":"Kiteis where robots get trained. Augmentation, training, evaluation, and deployment in one platform. Whoever owns this layer decides how the next decade of robots gets trained. We intend to be that team.Robots are about to be deployed in labs, factories, warehouses, and homes. The models are getting very smart. The tooling is not keeping up. Teams write their own pipelines, run evaluations manually, and lose most of their time to infrastructure instead of robot behavior.We are 2 founders. You would be the third person and the first founding engineer. The culture we are building: robot arms, 3D printers running, prototypes in various states of working. Watching a robot do the thing you meant it to do never gets old. Some days that happens here, some days at a customer site. Everyone helps and there is no ego in the room.The roleYou will decide how the platform is architected. Decide it, build it, and own it in production. This is hands on from day 1.Work you will pick up in the first 6 months:Training infrastructure: GPU orchestration, scheduling, resource allocation, memory management, cost per jobThe execution layer that runs augmentation, training, and evaluation jobs across cloud providersSimulator integrations and the dataset pipelineThe platform API and the surfaces on top of it: web, CLI, MCPTalking to the researchers and engineers using all of the above and fixing what they tell you is brokenWhat we are looking forAttributesWe filter on 3 things.Learning & execution speed.Robotics changes every week. A new dataset format, a new policy architecture, a new simulator. We need someone who reads the paper on Monday and has a prototype by Tuesday.Communication.Topics in robotics can get very abstract quickly, you will be in front of customers, on calls with their engineers, and you have to make hard ideas simple in real time.No ego. Weare willing to learn what we do not know yet to make what seems impossible possible. We are honest with each other, take feedback, and are optimists.Ideal ExperienceYou have architected a system from scratchGPU orchestration, scheduling, resourcing, memory. You know why a training job is slow and you know what it costsRobotics software, or infrastructure for robotics teamsYou have trained robots multiple times and are familiar with the spaceComfortable across the stack. Python, PyTorch, Docker, and enough front end to ship a feature end to endReasons not to take this jobIt is hard. Most of what we are building has no established answer in robotics yet. You will be working things out from first principles and some of them will be wrong.The field moves fast enough that work you did 3 months ago may need to be thrown away.Your role will be diverse. Infrastructure one week, a customer integration the next, front end after that.There is no schedule. There are customers, deadlines we set ourselves, and things that break.SF is our central office and we value in-person time. If you are based abroad we will arrange something personally, but you will still travel often.You will travel to customer sites. Robots are physical and you cannot debug a deployment over Slack.You will be client facing. You will sit 1 to 1 with engineers from other teams and work on their problems.What you getMeaningful equity. You are the first engineer and the number reflects that.Competitive cash compensation.Full ownership of the platform architecture, with no legacy system and no one to ask for permission.The work itself. Robotics training infrastructure is being defined right now and very few people are working on it. What you build here changes how the next generation of robotics teams trains autonomous robots.","datePosted":"2026-09-03T13:00:22.745Z","dateModified":"2026-09-03T13:00:22.745Z","hiringOrganization":{"@type":"Organization","name":"Kite Ml","sameAs":"https://jobsearcher.com"},"jobLocation":{"@type":"Place","address":{"@type":"PostalAddress","addressLocality":"Sonoma","addressRegion":"CA","addressCountry":"US"}},"identifier":{"@type":"PropertyValue","name":"JobSearcher","value":"d75761188c0fbe916d7cdcee"},"url":"https://jobsearcher.com/jobs/d75761188c0fbe916d7cdcee"}}