About the role
structured by ORI← All rolesMember of Technical Staff (Models)Train the models that make our AI Employees better at their jobs.Apply on Work at a StartupLocationSan FranciscoTeamEngineering, ModelsJob typeFull-timeExperience3+ years of experienceVisaUS citizen or visa holder onlyAbout JarminJarmin builds AI Employees for human-level…
What you will do
- Turning agent traces and expert feedback into training data
- Reward modeling and graders for non-verifiable outcomes
- Distillation into smaller, faster, cheaper models
- Designing and running training experiments
What they are looking for
- Have trained or fine-tuned models and shipped the result into a product
- Have hands-on experience with SFT, preference optimization, or RLHF
- Can read traces and tell whether a metric measures the thing that matters
- Have strong software engineering fundamentals alongside ML depth
- Want to work in person in San Francisco
Nice to have
- You've adapted open-weight models to a specialized domain
- You've built training data or evaluation infrastructure
- You contribute to open source projects
Before you apply
- Must be a US citizen or visa holder to apply.
Full posting text
← All rolesMember of Technical Staff (Models)Train the models that make our AI Employees better at their jobs.Apply on Work at a StartupLocationSan FranciscoTeamEngineering, ModelsJob typeFull-timeExperience3+ years of experienceVisaUS citizen or visa holder onlyAbout JarminJarmin builds AI Employees for human-level roles. An AI Employee learns its responsibilities, understands your data, embeds into your systems, and becomes part of your team's day-to-day workflow.
Companies hire AI ML Engineers, Account Executives, Paralegals, Financial Analysts, Compliance Analysts, and more.Our founding team is from Meta Superintelligence Labs (Staff Engineers), Apple (Machine Learning), AWS (scaling infrastructure), Lockheed Martin (Research), and JPMorganChase, with experience building AI and ML systems and automating the complex work of researchers, data scientists, and ML engineers.We are a Y Combinator Fall 2025 company in San Francisco.About the roleAs a Member of Technical Staff on Models, you'll own the post-training loop that improves the performance of our AI Employees.
You will be defining how we train models to do mission critical work in the real world.Areas you may work inPost-training and fine-tuningTurning agent traces and expert feedback into training dataReward modeling and graders for non-verifiable outcomesDistillation into smaller, faster, cheaper modelsDesigning and running training experimentsYou may be a good fit if youHave trained or fine-tuned models and shipped the result into a productHave hands-on experience with SFT, preference optimization, or RLHFCan read traces and tell whether a metric measures the thing that mattersHave strong software engineering fundamentals alongside ML depthWant to work in person in San FranciscoEven betterYou've adapted open-weight models to a specialized domainYou've built training data or evaluation infrastructureYou contribute to open source projectsHow we workOur core team is in person, resourceful, and building the future we want to live in.We give feedback early and often, so you always know where you stand and where you're headed.Interested?
Applications are handled on Work at a Startup. If you have a question before applying, write to careers@jarmin.ai.Apply on Work at a Startup