Senior Machine Learning Engineer - VLM
Torc RoboticsVisit Torc Robotics website
Torc Robotics is a Daimler Truck subsidiary developing Level 4 autonomous-driving software and systems for long-haul Freightliner Cascadia trucks.
Blacksburg, United States
About Torc Robotics
Founded in 2005 as Torc Technologies, Torc Robotics commercializes self-driving trucking technology, including its virtual-driver software, for freight operations.
Skills
ArrowAthenaAwsBevComputer VisionDatabricksDataset CurationDeep LearningDense CaptioningDepth EstimationDockerDynamodbEcsGithub ActionsLambdaLancedbLightningMcapMlflowMlopsObject DetectionOpenglPandasParquetPythonPytorchRayRosS3Semantic EmbeddingSemantic SegmentationSensor FusionSglangSparkStep FunctionsTerraformTrackingVlaVllmVlm
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will build cloud-based pipelines that turn logged sensor data into VLM and VLA training datasets. You will develop auto-labeling and reasoning-grounded annotations, curate long-tail data, define quality metrics, support model-training delivery, and mentor engineers.
Requirements
- Bachelor's degree plus 6+ years or master's degree plus 3+ years of relevant experience.
- Computer vision and deep learning experience, including at least two of object detection, tracking, sensor fusion, semantic segmentation, BEV, and depth estimation.
- Hands-on VLM, open-vocabulary recognition, zero-shot recognition, dense captioning, or semantic-embedding experience.
- Dataset curation and large-scale Parquet processing experience.
- Experience with PyTorch, Lightning, Ray, Spark, or equivalent frameworks.
- MLOps, experiment tracking, model registries, MLflow, Weights & Biases, and model evaluation experience.
- Strong Python development, cloud development environments, GitHub Actions, and Docker.
Responsibilities
- Design, implement, test, and deploy cloud-based VLM/VLA dataset pipelines.
- Develop VLM-assisted auto-labeling for detection, captioning, semantic enrichment, and scenario descriptions.
- Generate reasoning-grounded annotations aligned with ego-motion and trajectories.
- Mine rare and high-uncertainty scenarios and curate targeted datasets.
- Define dataset schemas, quality metrics, and validation processes.
- Route model failures into relabeling and retraining loops.
- Deliver datasets into training pipelines.
- Build distributed, reproducible data pipelines.
- Lead design reviews, establish coding and annotation standards, and mentor engineers.
Benefits
- Bonus component and stock options.
- 100% paid medical, dental, and vision premiums for full-time employees.
- 401K plan with a 6% employer match.
- Flexible schedule and generous paid vacation available immediately after start date.
- Company-wide holiday office closures.
- AD+D and life insurance.