Biogen Scientist I, Machine Learning in Cambridge, Massachusetts
About This Role * The Molecular Informatics team at Biogen is searching for a scientist interested in machine learning applications and transformative artificial intelligence algorithms to accelerate and improve the established small molecule design cycle that consists of idea generation, evaluation, and synthesis. This is a unique opportunity to join a cross-functional team working at the intersection of computational, chemical, and biological sciences and have direct access to a wealth of structured and unstructured data across the Biogen research organization to make a massive impact on redefining the early drug-discovery platform. *What You’ll Do * * Develop fully integrated molecular design generation workflows by applying deep reinforcement learning and other artificial intelligence algorithms in combinations with Open-Source cheminformatics toolkits. * Evaluate and apply current state-of-the-art transferable machine learning potentials for small molecules conformational energies based on deep learning algorithms. * Collaborate closely with the Computational Chemistry group and other cross-functional Drug Discovery teams to apply novel algorithms to relevant therapeutic projects. * Publish original work in top peer-reviewed journals as well as present his/her research work at internal and external meetings *Why Biogen? Our mission to find therapies for neurological and rare diseases is a unique focus within our industry and this shared purpose is what connects us as a team. We work together to overcome obstacles and to follow the science. Who You Are You are a creative and highly motivated scientist with demonstrated expertise in the field of machine learning and cheminformatics. You stay current with the scientific literature and apply the latest advances in computational methods to address drug discovery challenges. Thinks independently to solve complex problems. *Required Skills * * Ph.D. in a computational field (e.g. cheminformatics, computer science) * 1+ years postdoctoral experience * Track record of publications, and/or scientific presentations *Preferred Skills * * Good working experience with modern deep generative models for molecule generations, such as VAE, RNN, LSTM, and GAN. * Experience with multi-objective optimization algorithms, and their application to deep reinforcement learning architectures. * Experience building CNNs, RNNs, and other neural network infrastructure from scratch using deep learning libraries (e.g. TensorFlow, Keras, PyTorch). * Advanced knowledge of the Python programming language and related open-source libraries. * Machine learning experience building classical statistical models (e.g. Random Forests, SVM, Light GBM, XGBOOST) and basic statistics All your information will be kept confidential according to EEO guidelines.