Internship - AI Engineer - Multimodal Pre-trained Transformer Models for En

<b> Job title: </b><br> <br> Internship - AI Engineer - Document AI - Multimodal Pre-trained Transformer Models for Energy Industry (6 months)<br><br><b> About Us: </b><br> <br> We are a global technology company, driving energy innovation for a balanced planet.<br> <br> At SLB we create amazing technology that unlocks access to energy for the benefit of all. That is our purpose. As innovators, that's been our mission for 100 years. We are facing the world's greatest balancing act- how to simultaneously reduce emissions and meet the world's growing energy demands. We're working on that answer. Every day, a step closer.<br> <br> Our collective future depends on decarbonizing the fossil fuel industry, while innovating a new energy landscape. It's what drives us. Ensuring progress for people and the planet, on the journey to net zero and beyond. For a balanced planet.<br><br><b> Our Purpose </b><br> <br> Together, we create amazing technology that unlocks access to energy for the benefit of all.<br> You can find out more about us on company website<br><br><b> Location: </b> Clamart, France<br> <br> Come and Join SLB's AI Lab in Paris. We are currently offering internship to bright minds specialized in Data Science and Artificial Intelligence. Discover a multinational company. We have brought a little bit of the Silicon Valley in Paris. Experience working within a team of young and fun passionate Data Scientists, tackling real business challenges, in tandem with business experts who are sitting at your desk.<br> <br> The Artificial Intelligence & Machine Learning Data Scientist helps develop software and processes that can be used for robotics, artificial intelligence programs and application. In close collaboration with the business and metiers, the data scientist offers mathematical and statistical models from the collected data to augment, improve or speed up human decisions within the Energy sector. With SLB you will be given the opportunity to apply your expertise and deploy deep learning solutions at scale on real-world problems, supporting many areas of SLB's business. SLB is the first Energy service company to move its processes and workflows in the cloud. This gives the Embedded AI Lab the perfect opportunities to leverage these innovative technologies and resources. As part of the Embedded AI Lab you will be able to test, experiment and research with the bleeding edge environment with Petabits and Petabits of data. You will be in charge of applying research and delivery of Proof of Concepts solutions, responding to a clear and specific business needs..<br><br><b> Description: </b><br> <br> Document AI has shown great potential in downstream tasks such as information retrieval, document classification, question-answering, etc. Recently, multimodal pretrained document models [1, 2, 3] are proposed to jointly take multiple modalities such as textual, visual, and layout information into consideration, and bring breakthroughs in almost all document-related downstream tasks. <br> <br> However, most open-sourced pre-trained document models are built on public datasets in general domains. When dealing with domain-specific problems, domain experts are not satisfied with the performance of the pre-trained models, even after fine-tuning. Besides, not all domain-specific downstream tasks are handled by the current pre-trained document AI models, such as figure reconstruction, data restructuration, knowledge base construction, etc.<br> <br> As one of the key actors in the energy industry, SLB supports fundamental research in this domain. A domain-specific multimodal pretrain-finetuning document AI model for the energy industry will improve the efficiency of the relevant downstream tasks and knowledge management.<br> <br> Therefore, in this internship, we are looking to build a domain-specific multimodal document AI model for a selection of downstream tasks in the energy industry. Based on the previous work done by the AI Lab NLP team, the intern will have opportunities to explore the current SOTA models, extend them to more challenging tasks and directly see the impact on the business use cases.<br> <br> You will learn and practice:<br> <ul> <li> Multimodal pre-trained transformer models, like BERT, LayoutLM v2 and v3</li> <li> Applications of NLP algorithms, (cross-lingual) information extraction, rich text document analysis, etc..</li> </ul> <br><br><b> Required </b> <b>skills: </b> <br> <ul> <li> Applied mathematics, probability & statistics, deep learning</li> <li> Natural Language Processing (NLP)</li> <li> Programming language: Python, PyTorch or Keras/Tensorflow</li> </ul> <br><br><b> References </b> : <br> <ul> <li> [1] Huang, Yupan, et al.