This list contains only the countries for which job offers have been published in the selected language (e.g., in the French version, only job offers written in French are displayed, and in the English version, only those in English).
You’ll work directly alongside our Principal AI Engineer, who sets the technical direction for the system. Your job is to be an implementation and experimentation partner on the agentic system: the person who can take a research idea and turn it into a runnable experiment, interpret what the results mean, and help decide what to try next. As the system matures, the role will evolve toward owning the production architecture on AWS.
Job Responsibility:
Design and run experiments to evaluate what works: tool structures, reasoning patterns, retrieval approaches, optimization integrations. Interpret what the results are telling you about the system’s design
Build fast, readable prototypes that are easy to modify and discard when an experiment doesn’t pan out
Work iteratively with the team to figure out what the right agentic architecture is, not just how to implement a predetermined one
Diagnose failures clearly, distinguishing engineering problems from prompt design problems from domain modeling problems
Communicate findings clearly to a technically deep team that includes optical scientists and engineers
Grow into ownership of the production architecture on AWS as the system matures
Requirements:
Bachelor’s degree in physics, EE, or applied math
5+ years’ post-education work experience
Hands-on experience building agentic or LLM-integrated systems. You have debugged a broken tool loop, managed multi-turn state, and worked with at least one agentic framework (LangGraph, Strands Agents, Claude Agent, or similar)
Production-quality Python. Not notebook Python, but code that handles errors, is testable, and can be read by someone else six months from now
Meaningful AWS experience. You have built and operated real systems on AWS, understand core services (ECS, Lambda, S3, IAM, SQS, or related services), and have debugged things that broke in production
A demonstrated ability to design informative experiments. You know the difference between an experiment that answers a question and one that just generates activity
Comfort operating without a complete roadmap. You can make good decisions under ambiguity and know when to ask versus when to try
US Citizenship
Nice to have:
Master’s degree in physics, EE, or applied math
Familiarity with optical design concepts or software (e.g., Zemax, Code V)
Exposure to reinforcement learning concepts
Experience with MATLAB or numerical computing workflows
Background in advanced mathematics: optimization theory, numerical methods, linear algebra, or differential equations