Staff Engineer World Model Development

2 hours agoLeadSalary: 185K - 325KBostonOnsiteFull TimeAiJobs by Merlin

Merlin is a Boston-based aerospace and defense technology company developing AI-powered, aircraft-agnostic autonomy software for fixed-wing aircraft.

Boston, USA
About Merlin

Merlin’s Merlin Pilot combines hardware and software intended to perform takeoff-to-touchdown flight functions for military and civil aircraft, supporting reduced-crew and autonomous operations. The operating business began as Apollo Flight Research in 2018, became Merlin Labs in 2020, and is now operated under public parent Merlin, Inc.

View jobs by Merlin

Skills

Candidate Availability

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will design, train, and evaluate predictive world models for aircraft, environments, and traffic. You will build rollout systems for plan evaluation, deliver calibrated uncertainty and out-of-distribution signals, define model validity boundaries, address sim-to-real gaps, benchmark learned and physics-based models, and maintain reproducible evaluations.

Requirements

  • Degree in Computer Science, Artificial Intelligence, Data Science, Computer Engineering, Applied Math, or a related subject
  • 10+ years building and training deep learning models
  • Experience with sequence, dynamics, video, or trajectory prediction
  • Strong PyTorch skills
  • Knowledge of distributed training, mixed precision, experiment tracking, and training debugging
  • Knowledge of dynamical systems, state estimation, or control
  • Experience with evaluation and uncertainty quantification
  • Experience with real-world sensor and telemetry data

Responsibilities

  • Design, train, and evaluate world models for aircraft state, environment, and traffic
  • Build rollout machinery for evaluating candidate plans
  • Deliver calibrated predictive uncertainty and out-of-distribution signals
  • Characterize model validity boundaries across flight conditions
  • Train in simulation and validate against flight data
  • Feed simulator discrepancies back into simulator fidelity
  • Benchmark learned dynamics against physics-based and flight-controls models
  • Contribute to reproducible evaluation infrastructure

Benefits

  • Equity grants
  • Catered lunches
  • Snacks and beverages
  • Health insurance
  • Dental insurance
  • Life insurance
  • Unlimited vacation
  • 401k matching