Postdoc: AI-Driven Reactive Robotics for Isotope Production

Brookhaven National Laboratory; Upton (NY); On-site; USD 72,000 - 85,000; Full time

Job summary

Brookhaven National Laboratory in Upton, New York, seeks a postdoctoral researcher to develop an AI-driven robotic system for chemical purification in radioisotope production. The role emphasizes fast, reactive fine-motor control and real-time corrections, integrated with a vision-language planning model.

The successful candidate will work across robotics, radiochemistry, and computing, publish in peer‑reviewed venues, and contribute to unstructured lab environments.

Qualifications

• PhD in mechanical engineering, electrical engineering, computer science, physics, mathematics, statistics/data science, or closely related field.
• Backgrounds in robotics, controls, or machine learning are especially relevant.
• Strong programming skills in Python with ML frameworks (e.g. PyTorch).
• Peer‑reviewed publications in relevant fields are expected.

Responsibilities

• Design and build a dual-system robotic architecture with a vision-language model for task planning and a fast learned controller for reactive manipulation.
• Develop reactive control layer using suitable approach (e.g. MPC, learned motion priors, or latent models).
• Build the 3D perception stack for fluid-state estimation through glassware using pretrained encoders.
• Develop and validate on a cold bench before active material; publish results in peer‑reviewed venues.
• Collaborate across robotics, radiochemistry, and computing communities.

Skills

Robotics & controls

Machine learning

Python programming

PyTorch

Vision-language models

Education

PhD in a technical field

Tools

PyTorch

3D perception tools

Job description

Brookhaven National Laboratory in Upton, New York, seeks a postdoctoral researcher to develop an AI-driven robotic system for chemical purification in radioisotope production. The role emphasizes fast, reactive fine-motor control and real-time corrections, integrated with a vision-language planning model.

The successful candidate will work across robotics, radiochemistry, and computing, publish in peer‑reviewed venues, and contribute to unstructured lab environments.