Machine Learning Engineer, Applied

Trace; New York (NY); On-site; USD 185,000 - 245,000; Full time

Job summary

Trace Labs is seeking a capable ML engineer with a research mindset to bridge ML engineering and research. You will drive end-to-end model development across varied data types, from data collection to deployment in production.

You will collaborate with the CV team to ship models quickly, leveraging PyTorch or JAX and strong ML fundamentals. The role emphasizes fast delivery and the ability to adapt to multiple data modalities.

Qualifications

• 2+ years of hands-on experience training deep learning models on real-world data.
• Proficient in Python with modern DL frameworks (PyTorch or JAX).
• Solid ML fundamentals: architectures, optimization, loss design, evaluation.
• Comfort with messy real-world data such as video, time series, or sensor streams.
• An extraordinary bias toward shipping and ownership of problems.

Responsibilities

• Own models end to end: data, training, evaluation, deployment into production.
• Build training pipelines for multimodal datasets (video, sensor streams, language).
• Collaborate with Head of Engineering and CV team to deploy and improve models.
• Move quickly between diverse problems and build needed tooling.

Skills

Python

ML engineering

Strong shipping bias

Research mindset

Education

Bachelor's degree in Computer Science or related

Tools

PyTorch

JAX

Job description

About Trace

Trace Labs is building the data infrastructure for physical AI.

About Trace

Trace Labs is building the data infrastructure for physical AI.
Physical AI has the potential to transform how work gets done in the real world, from robotics to embodied systems that can see, move, and interact with their environment. But today, progress is held back by one big gap: there's no scalable way to collect high-quality, real-world training data. Frontier robotics models are trained on far less data than language models, because there's no "internet of robotics data." Trace exists to change that. We capture how humans actually interact with the physical world, at scale, and we build the ML systems that make every hour of that data more valuable.
We're an early, deeply technical team. We move fast, we care about quality, and we believe the teams with the best data will build the best robots.

The role

This role sits between ML engineering and research. You'll bring strong ML engineering skills, a research mindset, and deep learning fundamentals to the hardest problems we see coming at Trace, then ship the answers into production.
You'll work closely with our computer vision team, but this role is broader and more product-focused. You won't live on one type of data. One month you might be working with sensor signals, the next with hand tracking video, the next with language. What matters is getting a high-quality model working fast and putting it to use.
We care much more about what you've built and shipped than where you've published. If you're earlier in your career but have a clear track record of moving fast and owning hard problems, we want to talk.

What You Will Do

• Own models end to end: data, training, evaluation, and deployment into our annotation pipeline.
• Build training pipelines for large, multimodal datasets, including video, sensor streams, and language.
• Partner with our Head of Engineering and the CV team to get models into production and keep them improving.
• Start with the simplest approach that works, then make it better. Pull from recent papers when they help you ship, not as the end goal.
• Move quickly between very different problems, and build the tooling you need along the way.

What we're looking for

• A BS in Computer Science, Electrical Engineering, Mathematics, Aerospace, or a related field, or equivalent practical experience. An MS is a plus, not a must.
• 2+ years of industry of hands-on experience training deep learning models on real-world data, at a company, startup, or national lab. Earlier in your career is okay if your track record shows it.
• Strong proficiency in Python and a modern deep learning framework such as PyTorch or JAX.
• A solid grounding in ML fundamentals, including architectures, optimization, loss design, and evaluation.
• Comfort with messy, real-world data such as video, time series, or sensor streams.
• An extraordinary bias toward shipping. You go from idea to working model fast, and you make good calls under uncertainty.
• High agency. You find the problem, own it, and get it done without waiting to be told.
• Range. You're excited to work across different kinds of data and problems, not just one narrow specialty.

Why Trace Labs

• A foundational problem. The data layer for physical AI is still being defined. You'll help define it.
• Data at scale from day one. We have the resources to collect it. Your job is to make it count.
• Real ownership. Your models ship into production and directly shape what our data can do.
• Room to grow. Go deeper into ML, take on bigger systems, or both.

Compensation Range: $185K - $245K