Senior Computer Vision Engineer
hace 13 días
Bilbao
Overview Your role will be based in Vitoria-Gasteiz (Hybrid) to make key contributions to AI projects with a primary focus on computer vision, alongside natural language processing, foundational models and other general AI/ML solutions aligned with PEP AI roadmap within the global digital transformation journey within PepsiCo. You will be part of a collaborative interdisciplinary team around data and analytics, where you will explore new technologies, and design the adoption of new products in the domain of computer vision AI to support the development of digital solutions to create business decision making. Your work will span retail analytics, quality control systems, supply chain optimization, and consumer insights through visual intelligence. On the one hand, you will work closely with process owners, product owners and final business users. This will provide you the correct visibility and understanding of criticality of your developments to business stakeholders and S&T functional leaders. On the other hand, you will work closely with other data scientists to transfer your findings to ensure the scalability, robustness and reusability of the solutions that you develop. You will be an internal ambassador of the team's culture around data and analytics. You will provide stewardship to colleagues in the solution development landscape and further lead the onboarding of new developments in AI within the PepsiCo domain leading the development of best practices and standard ways of working. Responsibilities Active contributor to code development in projects and services related to computer vision, object detection, image segmentation, video analytics, machine learning, mathematical modeling, AI and ML/AI Operations, and deep learning. Design and deploy CV models for real-world applications including retail execution monitoring, manufacturing quality control, supply chain optimization, and visual search. Active contributor in innovation activities, exploring the boundaries of technology to create AI solutions with the ability to continually prove accuracy and with a design goal of productization/reusability. Create documentation for learnings and knowledge transfer. Create reusable packages or libraries. Partner with data engineers to ensure data access for discovery and proper data is prepared for model consumption. Partner with other data scientists working on industrialization. Occasionally, coordinate work activities with business teams, other services and as required. Communicate with data scientists and other S&T stakeholders in the process of service design, training and knowledge transfer. Support large-scale experimentation and build analytical models. Set KPIs and metrics to evaluate analytics solution given a particular use case. Refine requirements into modelling problems. Influence product teams through data-based recommendations. Research in state-of-the-art methodologies. Qualifications Master's Degree in Computer Science, Data Science, Mathematics, Statistics, Computer Vision, or a related quantitative field. Preferred: PhD in Computer Science, Machine Learning, Computer Vision, or related fields, with specialization in deep learning for visual recognition. Minimum: 4+ years of hands-on experience building, deploying, and evaluating computer vision models in a research or industry setting, with a strong focus on object detection, segmentation, or image classification applications. Ideal: 6+ years of experience, with demonstrable success in developing computer vision solutions, and delivering tangible business value. Computer Vision models and architectures. Experience with CNNs (ResNet, EfficientNet, MobileNet), Vision Transformers (ViT, Swin), object detection models (YOLO, Faster R-CNN, DETR), and segmentation models (U-Net, Mask R-CNN, SAM). Use cases: object detection, image classification, semantic/instance segmentation, visual quality inspection, OCR, video understanding, etc. Deep Learning Frameworks: Proficiency in PyTorch, TensorFlow, or similar frameworks. 8+ years working in Python, Java or other programming languages Python: Essential for data manipulation, model development, and integration. Data science libraries: NumPy, Pandas, Scikit-learn, Matplotlib. Computer Vision libraries: OpenCV, Pillow, Albumentations. Deep learning: PyTorch Vision, TorchVision, timm. Detection frameworks: Ultralytics, MMDetection, Detectron2. Experience with model optimization and deployment: ONNX, TensorRT, quantization, edge deployment. Data annotation tools: CVAT, Label Studio, or similar. Cloud Computing: Experience with platforms like AWS, Azure, and GCP. Azure DevOps: Knowledge of CI/CD pipelines and software development practices using Azure DevOps. Fluency in English: Strong written and verbal communication skills in English are essential for collaboration and conveying technical ideas. High proficiency in business storytelling and communicating insights in business consumable format. Experience with multimodal AI (vision-language models like CLIP, BLIP, GPT-4V) is a plus. Experience with Responsible AI and FAIR data is a plus. Experience with NLP and generative AI for language tasks is a plus. Track record of Developing AI products that have generated business value, and impact across multiple Industries. Strong problem-solving mindset and creatively to identify and address challenges and opportunities Strong written, oral communication, and interpersonal and leadership skills, ideally experience driving consensus Strong project management, organizational, and prioritization skills with the ability to deal with ambiguity while juggling multiple priorities Deep passion for AI technologies and their real-life applications Ability to think strategically within business context and work at a fast pace
