Machine Learning Engineer

Are you ready for the toughest AI challenges involving graph neural networks? We have developed a graph-based patent search engine that allows our users to find documents that already contain their concepts. We are in search of a Machine Learning Engineer who can bring innovative ideas to life with excellent software engineering. You will be dealing with the most central part of the product - the search algorithm. We've converted 100 million patent documents into 100 million knowledge graphs to simplify information for machine learning. This has been made possible thanks to the semi-structured text of patents. These documents also contain valuable metadata like inter-document citations, which serve as labels for supervised training. All of these resources present an opportunity to delve further into text understanding than even the LLMs from the tech-giants. The current model we use is the Graph Transformer (check out our blog about the graph transformer). Therefore, both the NLP transformers and the GNN models are useful. We have published a paper at the SIGIR 2023 conference about our overall approach (read our paper). The daily work is performed in a small mixed team. Remote work is acceptable, but we expect to meet at least every three months in person. There is also a separate team providing MLOps support. Technology stack: - PyTorch for search model development - ML services deployed as Python microservices - Google Cloud for computation Requirements: - Excellence in applied deep learning, ideally also with GNNs - A genuine willingness to learn - Knowledge about NLP - Solid social skills - Proficiency in English, which is our company language - Residing ideally in Finland, but at least you must be able to work in European time zones Benefits: - A rare opportunity to join an early-stage, disruptive company with a unique edge, ambitious plans, and a proven track record of success in a growing $2bn industry. - Opportunities for growth. We will help you excel in this role and develop skills to help you advance. You will be part of a highly skilled and supportive team. - A highly competitive financial package and other benefits, including six weeks of holiday. - In Finland, additional benefits include lunch, sports and culture perks, commuting benefits, extended occupational health care, including dental care, and tools of your choice. Next Steps: Our recruitment process includes: 1. An informal interview with our AI Team Lead Sebastian and Head of People and Talent Mari to discuss your background and ensure that our expectations align with yours. 2. We would like to see an example of your code. You can share your Github profile with us or complete a coding task. 3. A technical interview with Sebastian and an AI team member. 4. A final call with our CTO Juho. Does this sound appealing? Submit your application now! About IPRally: IPRally is revolutionizing the way patents are found and categorized. We have brought transparent and reliable AI applications to patent search, analysis, and classification processes, saving our customers time, resources, and money. We are transforming the complex and tedious process of finding crucial information for our customers, but more importantly, we help innovative companies to promptly find and exploit opportunities and significantly reduce risk. IPRally was established in Helsinki, Finland in 2018. We were created with the vision to automate processes that take away resources from inventors and patent professionals, helping them achieve their goals quicker. By modeling the technical content of patents and processing this through cutting-edge, proprietary graph neural networks, IPRally gives customers highly relevant information in seconds.