AI/ML Engineer for Grounded Science Systems

S27a; Cambridge (MA); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Health insurance

Lunch subsidy

401K with matching

Equity compensation

Job summary

Transfyr is building physical AI systems to capture real scientific work and turn it into actionable, machine-readable records. You will build end-to-end ML systems that learn from messy lab data and work closely with perception and software engineers to ground interpretations of actions and their impact.

You’ll design robust, scalable models that operate under partial observability and changing lab protocols, with emphasis on reliability and clear signaling for scientists and robotic actions.

Qualifications

• Strong programming expertise with software engineering and data systems.
• Deep understanding of machine learning from noisy, real-world data.
• Experience with multimodal data and grounding models in real workflows.
• Fluency in Python and modern ML frameworks (e.g., PyTorch).
• Ability to design data pipelines and scalable training infrastructure.
• Production awareness for moving prototypes to live systems.

Responsibilities

• Learn from messy real-world scientific execution data and build learnable representations.
• Design multimodal learning pipelines combining vision, sensor data, and metadata.
• Develop models that explain performance and handle uncertainty.
• Ensure models remain robust as protocols and environments change.
• Collaborate across ML, perception, software, and science teams to deploy in workflows.

Skills

Python

ML Fundamentals

Multimodal Learning

Data Pipelines

Production Systems

Systems Mindset

Software Engineering

Tools

PyTorch

Job description

Transfyr is building physical AI systems to capture real scientific work and turn it into actionable, machine-readable records. You will build end-to-end ML systems that learn from messy lab data and work closely with perception and software engineers to ground interpretations of actions and their impact.

You’ll design robust, scalable models that operate under partial observability and changing lab protocols, with emphasis on reliability and clear signaling for scientists and robotic actions.