Full-time

SWE, Research

San Francisco, California, United States · Remote

$250,000 - $370,000

SWE, ML Research

Most research teams are limited less by ideas than by how long it takes to test one. This team trains its own foundation models, already running in production at large scale, and their research group is trying new approaches rather than just scaling up. They need an engineer to make every experiment faster to run and easier to trust.

You’d sit inside the research group, working day to day with scientists and engineers rather than taking tickets from a separate platform team. Much of what’s needed doesn’t exist yet. You’d write the designs, choose the architecture and shape where research tooling goes next.

What you and the team will work on, please note we're looking for multiple engineers, so if you're stronger in a smaller amount of these areas, that's completely fine and we'd encourage you to apply :)

  • Scaling multi-node GPU training so the time from a new idea to a trained, scored model keeps shrinking

  • Data systems that take in huge volumes of messy multimodal data, version it and get it ready for training

  • Job scheduling, run tracking and storage across the research stack, working alongside the wider infrastructure team

  • Automated evaluation, so weak ideas are dropped early and strong ones move forward

  • Hardening promising research code into dependable systems that end up in shipped models

  • Simple internal apps that give non-technical teams a clear view of what’s happening in training

You’ll have hands-on experience with large-scale distributed training and high-volume data infrastructure, and know Kubernetes and ML workflow tooling well. You’ll have taken experimental code and turned it into something a whole team relies on. You’ll also be comfortable when priorities shift with results and some experiments fail. Speech or LLM infrastructure experience is a plus, but strong large-scale ML infra from any domain is relevant. They also want someone who builds their own AI tools and agents into how they work, not just uses them.

Typically looking for 5 to 12 years building software or ML systems.

📍 Remote across the US (SF or NYC office available for hybrid)
💰 Up to circa $340K base + bonus + stock
✅ Well funded unicorn | 35+ person research org

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Salary

$250,000 - $370,000

Location

San Francisco, California, United States · Remote

Contract

Full-time

Your consultant

Marc Powell

Full-time

SWE, Research

San Francisco, California, United States · Remote

$250,000 - $370,000

SWE, ML Research

Most research teams are limited less by ideas than by how long it takes to test one. This team trains its own foundation models, already running in production at large scale, and their research group is trying new approaches rather than just scaling up. They need an engineer to make every experiment faster to run and easier to trust.

You’d sit inside the research group, working day to day with scientists and engineers rather than taking tickets from a separate platform team. Much of what’s needed doesn’t exist yet. You’d write the designs, choose the architecture and shape where research tooling goes next.

What you and the team will work on, please note we're looking for multiple engineers, so if you're stronger in a smaller amount of these areas, that's completely fine and we'd encourage you to apply :)

  • Scaling multi-node GPU training so the time from a new idea to a trained, scored model keeps shrinking

  • Data systems that take in huge volumes of messy multimodal data, version it and get it ready for training

  • Job scheduling, run tracking and storage across the research stack, working alongside the wider infrastructure team

  • Automated evaluation, so weak ideas are dropped early and strong ones move forward

  • Hardening promising research code into dependable systems that end up in shipped models

  • Simple internal apps that give non-technical teams a clear view of what’s happening in training

You’ll have hands-on experience with large-scale distributed training and high-volume data infrastructure, and know Kubernetes and ML workflow tooling well. You’ll have taken experimental code and turned it into something a whole team relies on. You’ll also be comfortable when priorities shift with results and some experiments fail. Speech or LLM infrastructure experience is a plus, but strong large-scale ML infra from any domain is relevant. They also want someone who builds their own AI tools and agents into how they work, not just uses them.

Typically looking for 5 to 12 years building software or ML systems.

📍 Remote across the US (SF or NYC office available for hybrid)
💰 Up to circa $340K base + bonus + stock
✅ Well funded unicorn | 35+ person research org

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Your consultant

Marc Powell

Full-time

SWE, Research

San Francisco, California, United States · Remote

$250,000 - $370,000

SWE, ML Research

Most research teams are limited less by ideas than by how long it takes to test one. This team trains its own foundation models, already running in production at large scale, and their research group is trying new approaches rather than just scaling up. They need an engineer to make every experiment faster to run and easier to trust.

You’d sit inside the research group, working day to day with scientists and engineers rather than taking tickets from a separate platform team. Much of what’s needed doesn’t exist yet. You’d write the designs, choose the architecture and shape where research tooling goes next.

What you and the team will work on, please note we're looking for multiple engineers, so if you're stronger in a smaller amount of these areas, that's completely fine and we'd encourage you to apply :)

  • Scaling multi-node GPU training so the time from a new idea to a trained, scored model keeps shrinking

  • Data systems that take in huge volumes of messy multimodal data, version it and get it ready for training

  • Job scheduling, run tracking and storage across the research stack, working alongside the wider infrastructure team

  • Automated evaluation, so weak ideas are dropped early and strong ones move forward

  • Hardening promising research code into dependable systems that end up in shipped models

  • Simple internal apps that give non-technical teams a clear view of what’s happening in training

You’ll have hands-on experience with large-scale distributed training and high-volume data infrastructure, and know Kubernetes and ML workflow tooling well. You’ll have taken experimental code and turned it into something a whole team relies on. You’ll also be comfortable when priorities shift with results and some experiments fail. Speech or LLM infrastructure experience is a plus, but strong large-scale ML infra from any domain is relevant. They also want someone who builds their own AI tools and agents into how they work, not just uses them.

Typically looking for 5 to 12 years building software or ML systems.

📍 Remote across the US (SF or NYC office available for hybrid)
💰 Up to circa $340K base + bonus + stock
✅ Well funded unicorn | 35+ person research org

If it sounds relevant, we’d be happy to talk it through.

All applicants will receive a response.

Your consultant

Marc Powell

techire AI website footer image