Senior/Principal Data Scientist, Veeva

• 4+ years of experience as a data scientist (or 2+ years with a Ph.D. degree)
• Master's or Ph.D. in Computer Science, Artificial Intelligence, Computational Linguistics, or a related field
• Strong theoretical knowledge of Natural Language Processing, Machine Learning, and Deep Learning techniques
• Proven experience working with large language models and transformer architectures, such as GPT, BERT, or similar
• Familiarity with large-scale data processing and analysis, preferably within the medical domain
• Proficiency in Python and relevant NLP libraries (e.g., NLTK, SpaCy, Hugging Face Transformers)
• Experience in at least one framework for BigData (e.g., Ray, Spark) and one framework for Deep Learning (e.g., PyTorch, JAX)
• Experience working with cloud infrastructure (e.g., AWS, GCP, Azure) and containerization technologies (e.g., Docker, Kubernetes) and experience with bashing script
• Strong collaboration and communication skills, with the ability to work effectively in a cross-functional team
• Used to start-up environments
• Social competence and a team player
• High energy and ambitious
• Agile mindset

Desirable

• Your role will primarily involve developing LLM-based agents that are specialized in searching and extracting detailed information about Key Opinion Leaders (KOLs) in the healthcare sector
• You will craft an end-to-end human-in-the-loop pipeline to sift through a large array of unstructured medical documents—ranging from academic articles to clinical guidelines and meeting notes from therapeutic committees
• These agents will be equipped to perform semantic searches and provide precise answers to predefined queries concerning KOL-related data across various languages and disciplines
• Utilizing cloud infrastructure, you will build models capable of information extraction and question answering
• You will also collaborate with a dedicated team of software developers and DevOps engineers to refine these models and deploy them into production environments
• Adopt the latest technologies and trends in NLP to your platform
• Develop LLM-based agents capable of performing function calls and utilizing tools such as browsers for enhanced data interaction and retrieval
• Experience with Reinforcement Learning from Human Feedback (RLHF) methods such as Direct Preference Optimization (DPO) and Proximal Policy Optimization (PPO) for training LLMs based on human preferences
• Design, develop, and implement an end-to-end pipeline for extracting predefined categories of information from large-scale, unstructured data across multi-domain and multilingual settings
• Create a robust semantic search functionality that effectively answers user queries related to various aspects of the data
• Use and develop named entity recognition, entity-linking, slot-filling, few-shot learning, active learning, question/answering, dense passage retrieval, and other statistical techniques and models for information extraction and machine reading
• Deeply understand and analyze our data model per data source and geo-region and interpret model decisions
• Collaborate with data quality teams to define annotation task metrics, and perform qualitative and quantitative evaluation
• Utilize cloud infrastructure for model development, ensuring seamless collaboration with our team of software developers and DevOps engineers for efficient deployment to production