Senior Applied ML Engineer
for sections like About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule. Only include sections that have real specific content. We must not repeat job title as heading. Must not include Job ID etc. Remove boilerplate. Preserve all substantive info.
We need to parse the posting.
Key sections:
- Intro about CoreWeave (company description) maybe not needed as heading? Could be part of About the role? The instruction: Use for section headings (e.g. About the role, Responsibilities, Requirements, Qualifications, Skills, Benefits, Pay, Schedule). So we should create sections for those headings if content exists.
We have "About The Role" section: after "About The Role" heading in posting. Content: "You have trained LLMs to be SOTA on specific tasks. ...". That's prose.
Then "Required Qualifications" and "Preferred Qualifications". Also "Our Stack" maybe not a standard heading but could be considered Skills? Might be considered Skills. Also "Why CoreWeave?" includes culture, values, benefits. There's also "What We Offer" with benefits list. Also "The base salary range for this role is $182,000 to $242,000." That's Pay.
We need to decide which sections to include: About the role, Responsibilities? Not explicit responsibilities list; but the role description includes responsibilities: "You will be expected to generate and investigate research ideas towards solving the remaining obstacles to continuous learning in production. You will work with the broader CRAFT team to validate these research directions across real customer tasks." Also earlier: "You will have an opportunity to touch many parts of this stack." Could be responsibilities.
We can create a Responsibilities section with those bullet points.
Requirements: maybe combine Required Qualifications and Preferred Qualifications? The instruction says use headings Requirements, Qualifications, Skills. Could treat Required Qualifications as Requirements, Preferred Qualifications as Qualifications? Or combine.
Better: Use Requirements for required qualifications, Qualifications for preferred qualifications, Skills for technical stack (Kubernetes, Megatron, Temporal, Postgres, FastAPI, Python, PyTorch/JAX). Also maybe include "Strong programming skills in Python and hands on experience with modern ML frameworks such as PyTorch or JAX" as skill.
Pay: include salary range.
Benefits: include list from "What We Offer". Also mention discretionary bonus, equity awards, comprehensive benefits program.
Schedule: not mentioned; we should not include Schedule if not present.
We must not include placeholder text like "commensurate with experience". Not present.
We need to output only HTML fragment, no /. Use headings, - for lists,
for prose.
We must preserve all substantive info; do not invent or summarize away detail. So we need to include all details from the posting that are relevant.
Let's extract substantive content:
Company description: "CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com." This is maybe not needed as a heading but could be part of About the role? However the instruction says we should not repeat job title as heading; but we can include a paragraph about company as part of About the role? Might be okay. But we need to preserve all substantive info; we could include it in a
under About the role or maybe a separate section like "About CoreWeave"? Not in allowed headings list. Probably better to include it as part of About the role (since it's about the role's context). We'll include it as a paragraph.
Our Team description: "The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet. Our roadmap involves substantial technical risk; there are still major technical problems we’re facing without a proven solution. However, there is very little market risk. We’ve worked closely with the teams building agents at many of the top AI-native startups as well as large enterprises. If we can build this, everyone will want it. A self improving agent that learns from experience the way a human employee would could quickly capture a large fraction of the total inference market, which is worth tens of billions of dollars today and will be worth hundreds of billions in a few years."
That's substantive; could be part of About the role or maybe a separate section like "About the team". Not in allowed headings but we could still include as paragraph under About the role.
Now "About The Role" section content: after that heading.
Let's extract:
"You have trained LLMs to be SOTA on specific tasks. You have opinions on whether sequence-level or token-level importance ratios are more effective. You probably shared the ScaleRL paper in your group chats, and kicked off a few ablations after you read it. You will be expected to generate and investigate research ideas towards solving the remaining obstacles to continuous learning in production. You will work with the broader CRAFT team to validate these research directions across real customer tasks. We are very GPU rich and are ready to direct an enormous amount of compute at this effort. Beyond your role’s specific qualifications, we’re looking for strong engineers with great taste. The most important qualification by far is that you learn fast and can ship. This role will inevitably involve a lot of learning on the job; we’re building this airplane as we fly it. Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack. Although we operate as part of a larger company, the CRAFT team is small, has a large degree of autonomy and drives our own roadmap and priorities. This is an excellent role for someone looking to find their own company in the future."
That's prose.
Now Required Qualifications:
"Bachelor’s, Master’s, or PhD degree in Computer Science, Machine Learning, Robotics, or a related technical field
4+ years of experience in machine learning, or a PhD with 2+ years of relevant industry experience, with a strong focus on model training
Strong programming skills in Python and hands on experience with modern ML frameworks such as PyTorch or JAX
Strong understanding of LLM post-training techniques, including supervised fine-tuning, reinforcement learning, and on-policy distillation
Experience developing, evaluating, and deploying machine learning models in production environments
Strong research and problem solving skills, with the ability to work effectively in ambiguous, fast moving environments"
Preferred Qualifications:
"Publications, open source contributions, or demonstrated research impact in LLM post-training or agent learning
Experience with distributed training, GPU acceleration, and large scale model training systems
Experience leading technically complex projects or mentoring other engineers"
Our Stack:
"Kubernetes Megatron Temporal Postgres FastAPI"
Also earlier: "As Part Of Building Serverless RL, We Depend On The Following Libraries And Frameworks (among Many Others) Kubernetes Megatron Temporal Postgres FastAPI"
We could list those as skills.
Why CoreWeave? includes culture values: "Be Curious at Your Core Act Like an Owner Empower Employees Deliver Best-in-Class Client Experiences Achieve More Together". Also "We support and encourage an entrepreneurial outlook and independent thinking. We foster an environment that encourages collaboration and provides the opportunity to develop innovative solutions to complex problems. As we get set for takeoff, the growth opportunities within the organization are constantly expanding. You will be surrounded by some of the best talent in the industry, who will want to learn from you, too."
What We Offer: benefits list.
Also base salary range.
Now we need to decide sections:
- About the role: could include company description, team description, and the "About The Role" prose. Might be long but okay.
- Responsibilities: could extract from "You will be expected to generate and investigate research ideas..." and "You will work with the broader CRAFT team..." and maybe "Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack." That's more about opportunity but could be responsibilities.
- Requirements: Required Qualifications.
- Qualifications: Preferred Qualifications.
- Skills: Our Stack (Kubernetes, Megatron, Temporal, Postgres, FastAPI) plus maybe Python, PyTorch/JAX.
- Benefits: list from What We Offer plus discretionary bonus, equity awards, comprehensive benefits program.
- Pay: base salary range.
We must not include Schedule.
Now ensure we don't repeat content unnecessarily but preserve all substantive info. We'll need to include each piece somewhere.
Let's craft HTML.
We'll start with
About the role
then a containing company description, team description, and the role intro paragraphs. Might be long but okay.
Better to split into multiple
under same heading? The instruction: Use
for prose. Could have multiple
elements.
We'll do:
About the role
CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com.
The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet
headings, - for lists,
for prose.
We must preserve all substantive info; do not invent or summarize away detail. So we need to include all details from the posting that are relevant.
Let's extract substantive content:
Company description: "CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com." This is maybe not needed as a heading but could be part of About the role? However the instruction says we should not repeat job title as heading; but we can include a paragraph about company as part of About the role? Might be okay. But we need to preserve all substantive info; we could include it in a
under About the role or maybe a separate section like "About CoreWeave"? Not in allowed headings list. Probably better to include it as part of About the role (since it's about the role's context). We'll include it as a paragraph.
Our Team description: "The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet. Our roadmap involves substantial technical risk; there are still major technical problems we’re facing without a proven solution. However, there is very little market risk. We’ve worked closely with the teams building agents at many of the top AI-native startups as well as large enterprises. If we can build this, everyone will want it. A self improving agent that learns from experience the way a human employee would could quickly capture a large fraction of the total inference market, which is worth tens of billions of dollars today and will be worth hundreds of billions in a few years."
That's substantive; could be part of About the role or maybe a separate section like "About the team". Not in allowed headings but we could still include as paragraph under About the role.
Now "About The Role" section content: after that heading.
Let's extract:
"You have trained LLMs to be SOTA on specific tasks. You have opinions on whether sequence-level or token-level importance ratios are more effective. You probably shared the ScaleRL paper in your group chats, and kicked off a few ablations after you read it. You will be expected to generate and investigate research ideas towards solving the remaining obstacles to continuous learning in production. You will work with the broader CRAFT team to validate these research directions across real customer tasks. We are very GPU rich and are ready to direct an enormous amount of compute at this effort. Beyond your role’s specific qualifications, we’re looking for strong engineers with great taste. The most important qualification by far is that you learn fast and can ship. This role will inevitably involve a lot of learning on the job; we’re building this airplane as we fly it. Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack. Although we operate as part of a larger company, the CRAFT team is small, has a large degree of autonomy and drives our own roadmap and priorities. This is an excellent role for someone looking to find their own company in the future."
That's prose.
Now Required Qualifications:
"Bachelor’s, Master’s, or PhD degree in Computer Science, Machine Learning, Robotics, or a related technical field
4+ years of experience in machine learning, or a PhD with 2+ years of relevant industry experience, with a strong focus on model training
Strong programming skills in Python and hands on experience with modern ML frameworks such as PyTorch or JAX
Strong understanding of LLM post-training techniques, including supervised fine-tuning, reinforcement learning, and on-policy distillation
Experience developing, evaluating, and deploying machine learning models in production environments
Strong research and problem solving skills, with the ability to work effectively in ambiguous, fast moving environments"
Preferred Qualifications:
"Publications, open source contributions, or demonstrated research impact in LLM post-training or agent learning
Experience with distributed training, GPU acceleration, and large scale model training systems
Experience leading technically complex projects or mentoring other engineers"
Our Stack:
"Kubernetes Megatron Temporal Postgres FastAPI"
Also earlier: "As Part Of Building Serverless RL, We Depend On The Following Libraries And Frameworks (among Many Others) Kubernetes Megatron Temporal Postgres FastAPI"
We could list those as skills.
Why CoreWeave? includes culture values: "Be Curious at Your Core Act Like an Owner Empower Employees Deliver Best-in-Class Client Experiences Achieve More Together". Also "We support and encourage an entrepreneurial outlook and independent thinking. We foster an environment that encourages collaboration and provides the opportunity to develop innovative solutions to complex problems. As we get set for takeoff, the growth opportunities within the organization are constantly expanding. You will be surrounded by some of the best talent in the industry, who will want to learn from you, too."
What We Offer: benefits list.
Also base salary range.
Now we need to decide sections:
- About the role: could include company description, team description, and the "About The Role" prose. Might be long but okay.
- Responsibilities: could extract from "You will be expected to generate and investigate research ideas..." and "You will work with the broader CRAFT team..." and maybe "Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack." That's more about opportunity but could be responsibilities.
- Requirements: Required Qualifications.
- Qualifications: Preferred Qualifications.
- Skills: Our Stack (Kubernetes, Megatron, Temporal, Postgres, FastAPI) plus maybe Python, PyTorch/JAX.
- Benefits: list from What We Offer plus discretionary bonus, equity awards, comprehensive benefits program.
- Pay: base salary range.
We must not include Schedule.
Now ensure we don't repeat content unnecessarily but preserve all substantive info. We'll need to include each piece somewhere.
Let's craft HTML.
We'll start with
About the role
then a containing company description, team description, and the role intro paragraphs. Might be long but okay.
Better to split into multiple
under same heading? The instruction: Use
for prose. Could have multiple
elements.
We'll do:
About the role
CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com.
The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet
for prose. We must preserve all substantive info; do not invent or summarize away detail. So we need to include all details from the posting that are relevant. Let's extract substantive content: Company description: "CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com." This is maybe not needed as a heading but could be part of About the role? However the instruction says we should not repeat job title as heading; but we can include a paragraph about company as part of About the role? Might be okay. But we need to preserve all substantive info; we could include it in a
under About the role or maybe a separate section like "About CoreWeave"? Not in allowed headings list. Probably better to include it as part of About the role (since it's about the role's context). We'll include it as a paragraph. Our Team description: "The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet. Our roadmap involves substantial technical risk; there are still major technical problems we’re facing without a proven solution. However, there is very little market risk. We’ve worked closely with the teams building agents at many of the top AI-native startups as well as large enterprises. If we can build this, everyone will want it. A self improving agent that learns from experience the way a human employee would could quickly capture a large fraction of the total inference market, which is worth tens of billions of dollars today and will be worth hundreds of billions in a few years." That's substantive; could be part of About the role or maybe a separate section like "About the team". Not in allowed headings but we could still include as paragraph under About the role. Now "About The Role" section content: after that heading. Let's extract: "You have trained LLMs to be SOTA on specific tasks. You have opinions on whether sequence-level or token-level importance ratios are more effective. You probably shared the ScaleRL paper in your group chats, and kicked off a few ablations after you read it. You will be expected to generate and investigate research ideas towards solving the remaining obstacles to continuous learning in production. You will work with the broader CRAFT team to validate these research directions across real customer tasks. We are very GPU rich and are ready to direct an enormous amount of compute at this effort. Beyond your role’s specific qualifications, we’re looking for strong engineers with great taste. The most important qualification by far is that you learn fast and can ship. This role will inevitably involve a lot of learning on the job; we’re building this airplane as we fly it. Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack. Although we operate as part of a larger company, the CRAFT team is small, has a large degree of autonomy and drives our own roadmap and priorities. This is an excellent role for someone looking to find their own company in the future." That's prose. Now Required Qualifications: "Bachelor’s, Master’s, or PhD degree in Computer Science, Machine Learning, Robotics, or a related technical field 4+ years of experience in machine learning, or a PhD with 2+ years of relevant industry experience, with a strong focus on model training Strong programming skills in Python and hands on experience with modern ML frameworks such as PyTorch or JAX Strong understanding of LLM post-training techniques, including supervised fine-tuning, reinforcement learning, and on-policy distillation Experience developing, evaluating, and deploying machine learning models in production environments Strong research and problem solving skills, with the ability to work effectively in ambiguous, fast moving environments" Preferred Qualifications: "Publications, open source contributions, or demonstrated research impact in LLM post-training or agent learning Experience with distributed training, GPU acceleration, and large scale model training systems Experience leading technically complex projects or mentoring other engineers" Our Stack: "Kubernetes Megatron Temporal Postgres FastAPI" Also earlier: "As Part Of Building Serverless RL, We Depend On The Following Libraries And Frameworks (among Many Others) Kubernetes Megatron Temporal Postgres FastAPI" We could list those as skills. Why CoreWeave? includes culture values: "Be Curious at Your Core Act Like an Owner Empower Employees Deliver Best-in-Class Client Experiences Achieve More Together". Also "We support and encourage an entrepreneurial outlook and independent thinking. We foster an environment that encourages collaboration and provides the opportunity to develop innovative solutions to complex problems. As we get set for takeoff, the growth opportunities within the organization are constantly expanding. You will be surrounded by some of the best talent in the industry, who will want to learn from you, too." What We Offer: benefits list. Also base salary range. Now we need to decide sections: - About the role: could include company description, team description, and the "About The Role" prose. Might be long but okay. - Responsibilities: could extract from "You will be expected to generate and investigate research ideas..." and "You will work with the broader CRAFT team..." and maybe "Engineers on our team touch everything from CUDA kernels to high-performance LLM tracing dashboards, and you will have an opportunity to touch many parts of this stack." That's more about opportunity but could be responsibilities. - Requirements: Required Qualifications. - Qualifications: Preferred Qualifications. - Skills: Our Stack (Kubernetes, Megatron, Temporal, Postgres, FastAPI) plus maybe Python, PyTorch/JAX. - Benefits: list from What We Offer plus discretionary bonus, equity awards, comprehensive benefits program. - Pay: base salary range. We must not include Schedule. Now ensure we don't repeat content unnecessarily but preserve all substantive info. We'll need to include each piece somewhere. Let's craft HTML. We'll start with
About the role
then acontaining company description, team description, and the role intro paragraphs. Might be long but okay. Better to split into multiple
under same heading? The instruction: Use
for prose. Could have multiple
elements. We'll do:
About the role
CoreWeave is The Essential Cloud for AI™. Built for pioneers by pioneers, CoreWeave delivers a platform of technology, tools, and teams that enables innovators to build and scale AI with confidence. Trusted by leading AI labs, startups, and global enterprises, CoreWeave combines superior infrastructure performance with deep technical expertise to accelerate breakthroughs and turn compute into capability. Founded in 2017, CoreWeave became a publicly traded company (Nasdaq: CRWV) in March 2025. Learn more at www.coreweave.com.
The CRAFT (Capabilities, Routing, Adaptation & Fine-Tuning) Team At CoreWeave Is Building Tools To Help Agents Learn From Experience. This Is a Critical Step To Make Agents Reliable Enough To Perform Long Tasks Autonomously, In The Same Way Human Employees Are. We’re Systematically Identifying And Solving The Major Bottlenecks Between Today’s Tech And Those Future Self-improving Agents. So Far, We’ve Released ART, the easiest library for getting started with RL. Developed RULER, a general-purpose reward function that works across many diverse tasks. Built Serverless RL, an elegant API that gives RL practitioners full control over their data, environment, and reward function while letting them outsource the headaches of managing GPU infrastructure. These releases have a theme: we’re systematically tackling each major roadblock to successfully training self-improving agents. Several serious challenges remain. Building simulated environments often requires substantial human labor, and existing training methods are not data-efficient enough. We're laser-focused on solving these problems and making self-improvement a reality for agent developers. In startup terms, this is a classic hard-tech bet