Description
Merlin is hiring a World Model Engineer to design, train, evaluate, and deploy learned predictive models of aircraft, environments, and other traffic. The role owns the rollout machinery that generates candidate futures, handles predictive uncertainty and out-of-distribution signals, defines model validity boundaries, bridges simulation and real flight data, and benchmarks learned dynamics against physics-based models. Candidates need a relevant degree, 10+ years of deep learning experience, strong PyTorch skills, and expertise in dynamics, state estimation, control, evaluation, and uncertainty quantification. Remote applicants are welcome, with a preference for Boston-based candidates for a hybrid schedule; visa sponsorship is not available.
