Generative AI Lead Scientist
Evaluate Group, United Kingdom
Job Description
#LIHybrid
ABOUT NORSTELLA
Norstella is a group of prominent pharmaceutical solutions providers Evaluate, MMIT, Panalgo, The Dedham Group, Citeline that help clients navigate complexities at each step of the drug development life cycle, from pipeline to patient. For more information, please visit Norstella.com.
Evaluate is a global company providing outstanding market intelligence services for the Pharmaceutical, Medical Device, Financial and Consulting sectors, through the Evaluate Pharma, Evaluate Medtech, Evaluate Omnium and Evaluate Vantage online brands. Our international clients in Pharma and Biotech, Medtech, Banking and Consultancy regard Evaluate Pharma as the industry’s gold standard for timely and accurate analysis of reported drug sales, consensus sales forecasts, Ramp;D pipeline, markets and comprehensive company financials.
THE TEAM
In this role as a Generative AILead Scientist you will report into the Head of Data Science within Norstella. Your role will be to work with a dedicate group of data scientists and developers to use LLMs and related technologies to design and build products which will help answer complex customer questions directly, with citations and sources. We have a multi-functional team consisting of pharmaceutical industry experts, Ramp;D, data engineering and data scientists to rapidly prototype using both our existing and newly acquired datasets across the pharmaceutical life cycle.
SCOPE OF THE ROLE
In this role as a Generative AI Lead Scientist you will:
- Work with our Product leadership within Norstella to define and shape new offerings using Generative AI
- Manage a team of data scientists and developers to provide Generative AI-enabled microservices to the front end dev team
- Manage a roadmap of Generative AI work, mixing new product development and enhancements to existing services
- Define reference architectures suitable to answer complex questions, including via code interpreting, LLM tool use and leveraging secondary data science models
- Coach and train a team of data scientists and developers to use these architectures
- Stay up-to-date, constantly learning about advances in the field, and deliver periodic presentations to the data science team on these developments
HOW YOU’LL SUCCEED
Ultimately our goal is to smooth patient access to life-saving therapies. You will work with Ramp;D pharma specialists to understand a problem which is hindering developing and releasing effective new pharma products which we believe we can help with. After understanding the problem you will conceptualise potential solutions; this will involve breaking down the problem into individual steps, and identifying which of them can be solved with available LLMs and other algorithms. Finally, an overall solution can be packaged together with multiple models chained together, mixed with classical logic and business rules.
After conceiving potential solution(s), you will define a set of projects for the data scientists and python developers to implement your solution as a proof-of-concept. You will deliver indicative results from your PoC into datasets for exploration by the broader multi-functional team. You will also perform code reviews with the data science team to examine their implementation and consider ways of strengthening the final codebase and methodology.
After iterating the design with the multi-functional team as part of customer-led product development, you might convert your prototype into a full product. This will involve productionising code from you and the team to a high standard, containerisation, and deployment of the algorithm, usually as an API in AWS SageMaker, or for smaller pieces Lambdas which invoke models via API. Over time you may revisit this product, re-evaluate its performance, and redesign/improve as required.
Requirements
Essential Requirements
- Degree at Masters level or better in a STEM field such as Maths, Physics, Computer Science, Engineering, or equivalent practical experience
- Leadership experience, line managing data scientists or other highly specialised professionals
- Excellent communication skills when working with a broader development and product team
- Knowledge of Scrum, Kanban, and other Agile methodologies, as well as breaking down tasks into Jira tickets
- Proven experience using LLMs and prompt engineering inside applications to solve problems not tractable with older techniques
- Excellent knowledge of python and core data science libraries including Pandas, NumPy and other similar libraries
Nice to have
- Experience with AWS espec