Responsibilities
- Investigate and apply generative AI methods to produce realistic, varied synthetic image datasets for industrial use.
- Study diffusion models, GANs, image-to-image translation, and controllable generation to overcome data scarcity and class imbalance.
- Design and assess deep learning models for industrial visual inspection, including classification, defect detection, and segmentation.
- Implement convolutional and transformer-based neural networks, as well as multimodal system architectures.
- Apply transfer learning, self-supervised learning, few-shot learning, domain adaptation, and active learning techniques to reduce labeling needs.
- Enhance model performance in low-data environments using synthetic data augmentation strategies.
- Create systems to continuously monitor deployed vision models and data pipelines in production.
- Identify data drift, shifts in image quality, model degradation, confidence changes, and anomalous defect patterns.
- Detect out-of-distribution inputs and evolving operational conditions affecting model reliability.
- Formulate testable research questions, design experiments and ablation studies, and establish benchmarks for evaluation.
- Analyze results with scientific rigor and validate findings across multiple scenarios.
- Benchmark current methods and select those best suited for real-world industrial deployment.
- Develop proof-of-concept implementations and research prototypes using Python and modern deep learning frameworks.
- Partner with software, cloud, MLOps, manufacturing, and domain experts to integrate AI solutions into scalable platforms.
- Support full lifecycle development including data curation, training, validation, deployment, monitoring, and retraining.
Compensation
Not specified
Work Arrangement
Not specified
Team
Collaborative environment with global research, engineering, manufacturing, and product teams focused on industrial AI applications.
Contributions to Innovation
- Document research outcomes, present findings to technical and business audiences, and support publications, invention disclosures, and patent filings.
- Develop reusable frameworks and contribute to the organization’s intellectual property portfolio.
Cross-Functional Collaboration
- Work with global teams to understand inspection needs and translate practical challenges into defined AI research problems.
- Bridge research and production by converting prototypes into scalable, deployable AI components.
Not specified
