AV Simulation Domain Expert (Sr. Principal)

Role Description

HERE Technologies sits at a unique intersection: we own some of the world's most detailed map and drive data, and we are building the generative AI capabilities to turn that spatial intelligence into controllable, high-quality synthetic driving worlds. We are looking for a rare hybrid profile — someone who combines deep learning expertise in world foundation models, generative video, and transformers with hands-on AV simulation experience.

You understand both how to train and adapt large generative models (think Cosmos, Cosmos-Transfer, diffusion-based video models, latent world models) and how to ground them in map data and scenario semantics so the output is actually useful for training and validating perception and planning stacks. This is not a pure simulation role, and it is not a pure ML research role. It is the bridge between the two — and that bridge is where HERE's differentiation lives.

What you will do

• Drive the technical direction for map-grounded world foundation models.
• Train, fine-tune, and adapt generative models for driving scenario generation.
• Evaluate and extend state-of-the-art foundation models.
• Own the full ML lifecycle end-to-end.
• Lead proof-of-concept initiatives demonstrating map-grounded synthetic scenario generation.
• Define measurable success criteria focusing on ML training data utility.
• Deliver POC outcomes with clear recommendations backed by quantitative evidence.
• Bridge generative world models with classical simulation stacks.
• Author and programmatically generate OpenSCENARIO / OpenDRIVE definitions.
• Drive sim-to-real strategy: measure domain gap and define acceptable thresholds.
• Define what "good enough" synthetic data means for AV perception and planning.
• Establish validation frameworks combining objective metrics with expert evaluation protocols.
• Specify sensor fidelity requirements and how generative models should reproduce them.
• Interface with ML research teams on generative model architecture.
• Collaborate with perception and planning teams to ensure synthetic data improves performance.
• Translate business requirements into technical feasibility assessments.

Qualifications

• Proven experience training deep learning models end-to-end.
• Expertise in generative video, world models, or related generative AI research/engineering.
• Deep working knowledge of diffusion models, latent video models, and/or transformer-based world models.
• Experience with high-dimensional temporal or spatio-temporal data.
• Strong Python and PyTorch engineering fundamentals.
• Demonstrated ability to take ML models from research into production.
• 5+ years combined experience spanning AV simulation, perception/ML for AVs, or robotics simulation.
• Hands-on experience with at least one major simulation platform.
• Fluency with OpenDRIVE and OpenSCENARIO.
• Understanding of AV testing workflows.
• Understanding of what scenarios stress-test AV perception and planning systems.
• Ability to evaluate synthetic data for ML training utility.
• Experience with synthetic-to-real transfer or domain adaptation.
• Clear point of view on trade-offs between photorealism, label accuracy, controllability, and computational efficiency.

Nice-to-Have

• Hands-on experience with NVIDIA Cosmos, Cosmos-Transfer, or comparable world foundation models.
• Reinforcement learning experience.
• Experience with end-to-end driving models.
• Automotive, OEM, or other safety-critical ML deployment experience.
• Strong publication record in generative models or AV ML.
• Game engine experience for rendering and sensor simulation pipelines.
• Experience with PyTorch Lightning or similar large-scale training infrastructure.

Personal Attributes

• Bridge-builder: fluent translator between ML researchers, simulation engineers, AV domain experts, and product managers.
• Hands-on: you validate assumptions by training models and running simulations.
• Quality-obsessed: you define objective standards.
• Pragmatic: you balance "state-of-the-art realism" against "measurably useful for training."
• Systems thinker: you understand how every choice in data generation propagates into downstream model performance.