Principal Machine Learning Researcher (Physical AI)
About this role
Freeform builds AI-native manufacturing systems that unify software, hardware, and physics to produce industrial-scale parts at the speed of human ideation. By treating manufacturing as a single integrated system, the architecture enables the continuous generation of petabyte-scale, high-fidelity data that captures the physics of metal printing.
The Principal Machine Learning Researcher leads the development of advanced learning and control problems in this production-scale, AI-native metal manufacturing environment. Day-to-day work centers on creating machine learning methods that integrate large-scale physical data with physics-based simulation, embedding these models directly into closed-loop control and autonomy frameworks.
Ideal candidates bring a deep academic background in a quantitative discipline combined with extensive industry experience in machine learning applied to physical systems. Someone who thrives when bridging theory, experimentation, and production deployment will find this role a strong match for their capabilities.
What you will do
- Design and develop advanced machine learning models for complex, multi-physics manufacturing processes.
- Build hybrid modeling approaches that effectively combine first-principles physics with data-driven learning.
- Formulate learning-based models for prediction and control in production-scale metal additive manufacturing systems.
- Process and learn from large-scale, high-dimensional in-situ sensor data collected during printing.
- Implement unsupervised and self-supervised techniques to correlate process signals with part quality and geometry.
- Integrate learned models with physics-based simulation and digital twin frameworks.
- Contribute to real-time closed-loop control and autonomy systems operating on production hardware.
- Guide the integration of machine learning models into production software and manufacturing workflows.
What you need
- Hold a PhD in computer science, applied mathematics, physics, robotics, controls, or a closely related discipline.
- Demonstrate at least 5 years of experience in machine learning, applied research, or a related technical field.
- Maintain strong foundations in machine learning applied to physical systems, modeling, or control.
- Show proficiency in Python alongside at least one systems-level programming language such as C or C++.
- Bring practical experience working with large-scale, noisy, real-world datasets.
Nice to have
- 10+ years of industry experience
- Experience with hybrid physics–ML models, digital twins, or simulation-in-the-loop learning
- Background in autonomy, robotics, model predictive control, optimal control, or reinforcement learning for physical systems
- Experience with image-based or sensor-based inference in industrial or scientific settings
- Familiarity with computational geometry or geometric modeling
Benefits
- Significant stock option packages
- 100% employer-paid Medical, Dental, and Vision insurance
- Life insurance and Traditional or Roth 401(k)
- Relocation assistance provided
- Paid vacation, sick leave, company holidays, and parental leave
- Free daily catered lunch and dinner with fully stocked kitchenettes
Skills and tech stack
Questions candidates ask
What are the location and remote work policies for this role?
This role requires full-time onsite presence five days a week at the Hawthorne facility with very limited exceptions. The vertically integrated facility brings technology development, R&D, and production together under one roof.
What is the compensation range for the Principal Machine Learning Researcher?
The estimated salary range for this role is $200,000 to $500,000, complemented by a significant stock option package. Compensation is determined based on years of experience, educational background, and unique skills demonstrated during the interview process.
What kind of data will I be working with?
You will work with petabyte-scale, high-fidelity data capturing the physics of metal printing. This includes in-situ process signals, machine state data, geometry, and material outcomes.