Vision-Language Models and Generative AI (GenAI)

Roles & Responsibilities:

Conduct deep research in:

• Vision-Language and Multimodal AI for perception and semantic grounding
• Cross-modal representation learning for real-world sensor fusion (camera, lidar, radar, text)
• Multimodal generative models for scene prediction, intent inference, or simulation
• Efficient model architectures for edge deployment in automotive and factory systems
• Evaluation methods for explain ability, alignment, and safety of VLMs in mission-critical applications
• Spin newer research directions and drive AI research programs for autonomous driving, ADAS, and Industry 4.0 applications.
• Create new collaborations within and outside of Bosch in relevant domains.
• Contribute to Bosch's internal knowledge base, open research assets, and patent portfolio.
• Lead internal research clusters or thematic initiatives across autonomous systems or industrial AI.
• Mentor and guide research associates, interns, and young scientists.

Qualifications

Educational qualification:

Ph.D. in Computer Science / Machine Learning / AI / Computer Vision or equivalent

Experience:

8+ years (post PhD) in AI related to Vision and Language modalities, excellent exposure and hands on research in GenAI, VLMs, Multimodal AI, or Applied AI Research.

Mandatory/requires Skills:

Deep expertise in:

• Vision-Language Models (CLIP, Flamingo, Kosmos, BLIP, GIT) and multimodal transformers
• Open- and closed-source LLMs (e.g., LLaMA, GPT, Claude, Gemini) with visual grounding extensions
• Contrastive learning, cross-modal fusion, and structured generative outputs (e.g., scene graphs)
• PyTorch, HuggingFace, OpenCLIP, and deep learning stack for computer vision
• Evaluation on ADAS/mobility benchmarks (e.g., nuScenes, BDD100k) and industrial datasets
• Strong track record of publications in relevant AI/ML/vision venues
• Demonstrated capability to lead independent research programs
• Familiarity with multi-agent architectures, RLHF, and goal-conditioned VLMs for autonomous agents

Preferred Skills:

Hands-on experience with:

• Perception stacks for ADAS, SLAM, or autonomous robots
• Vision pipeline tools (MMDetection, Detectron2, YOLOv8) and video understanding models
• Semantic segmentation, depth estimation, 3D vision, and temporal models
• Industrial datasets and tasks: defect detection, visual inspection, operator assistance
• Lightweight or compressed VLMs for embedded hardware (e.g., in vehicle ECUs or factory edge)
• Knowledge of reinforcement learning or planning in embodied AI context
• Strong academic or industry research collaborations
• Understanding of Bosch domains and workflows in mobility and manufacturing