Research Engineer - Generative AI and Computer Vision

About the role

Bosch Global Software Technologies Private Limited, a 100% owned subsidiary of Robert Bosch GmbH and the largest software development centre of Bosch outside Germany, is hiring a Research Engineer in Bengaluru. The role advances state-of-the-art Generative AI and deep learning methods for real-world computer vision applications, combining applied research, experimentation and engineering to build robust AI solutions for image understanding, synthetic data generation, visual inspection and continuous model monitoring.

Key responsibilities

• Research and implement Generative AI approaches for synthetic image datasets, including diffusion models, GANs and image-to-image generation, to handle limited data, rare defects and class imbalance.
• Develop and evaluate deep learning models for industrial image inspection (classification, defect detection, anomaly detection, object detection, segmentation).
• Explore data-efficient methods such as transfer learning, self-supervised learning, few-shot learning and domain adaptation.
• Build methods to monitor deployed models and pipelines for data drift, performance degradation and out-of-distribution inputs.
• Design experiments and ablation studies, set benchmarks and compare state-of-the-art methods.
• Build proof-of-concept prototypes in Python and integrate them with software, cloud, MLOps, manufacturing and domain teams.
• Document findings and contribute to publications, patents and Bosch's intellectual property.

Required profile

• Ph.D., M.S. or M.Tech from a top Indian or foreign institute in CS, AI, ML, EE, ECE, robotics or a related field.
• At least 3 years of professional or applied research experience in computer vision, deep learning or Generative AI.
• Hands-on experience with CNNs, transformers, synthetic data pipelines and model evaluation metrics.
• Ability to work with global, interdisciplinary teams.

Required skills

• Python and PyTorch; NumPy, pandas, OpenCV, scikit-learn
• Vision Transformers and other transformer-based vision models
• Diffusion models, GANs, VAEs and image-to-image generation
• Hugging Face ecosystem
• Precision, recall, F1-score, AUC, ROC and model monitoring concepts
• Version control, testing, documentation and reproducible experimentation

Preferred: industrial quality inspection experience, synthetic defect generation, vision-language models, MLOps, Docker/Kubernetes, TensorRT/ONNX Runtime, and publications or patents.