Principal Scientist Data Science

Johnson & Johnson Johnson & Johnson · Pharma · Cambridge, MA +6

Principal Scientist role focused on developing and implementing ML, optimization, and GenAI solutions for clinical trial operations at Johnson & Johnson. The role involves building predictive models, optimization engines, and adapting LLMs for tasks like information extraction, protocol analysis, and schedule optimization. Requires a Ph.D. in a quantitative discipline and 5+ years of industry experience with ML, optimization, NLP, and GenAI, along with MLOps and programming skills.

What you'd actually do

  1. Conceive, develop, and implement ML, multi-objective optimization, GenAI solutions to support clinical trial operations.
  2. Leverage operational, RWD, and cost data to build ML predictive models and optimization engines to 1) predict outcomes of interest, 2) highlight tradeoffs between competing objectives, 3) recommend optimal operational scenarios, and 4) generate actionable insights, enabling early intervention and risk management.
  3. Adapt large language models (LLMs) for tailored information extraction and to create solutions including - conducting comparative analytics on clinical trial protocols and trial similarity assessment, - clinical trial data harmonization and standardization, - schedule of activity optimization, and - eligibility criteria evaluation,
  4. Stochastic enrollment simulations to forecast operational and patient journey outcomes including enrollment and study completion.
  5. Clearly articulate highly technical methods and results to diverse audiences and partners to drive decision-making.

Skills

Required

  • Ph.D. degree in a quantitative discipline (e.g., computer science, electrical and computer engineering, biostatistics, health economics, biomedical informatics, applied mathematics, or similar)
  • 5+ years of industry experience delivering on data science projects using ML predictive modeling, multi-objective optimization, natural language processing, and GenAI.
  • Hands-on experience with multi-modal ML predictive modeling and stochastic simulations for time-series forecasting.
  • Experience building multi-objective optimization engines to navigate complex trade-offs using evolutionary algorithms, reinforcement learning, or mixed-integer linear programming.
  • Experience with GenAI and clinical LLMs for document parsing and clinical concept disambiguation and harmonization.
  • Proficient in MLOps practices and tools (MLflow, Kedro); Git usage, CI/CD stacks (Jenkins, GitLab) DevOps tools.
  • Proficiency with programming languages Python and SQL
  • Experience with python LLM tools (e.g., DSPy, LangChain), optimization tools (e.g., pymoo) and ML tools (e.g., Scikit-learn, XGBoost, Optuna, PyMc)
  • Demonstrated experience and familiarity with clinical operational data, real world data, electronic health records and claims, and financial data.

Nice to have

  • Prior experience in a data science AI/ML role in healthcare, MedTech, and pharmaceutical industries.
  • Hand-on experience utilizing clinical trial protocols, registry, cost data, CTMS, EDC and/or EHR to build ML models for estimating operational outcomes or RWE outcomes.
  • Advanced Analytics
  • Coaching
  • Critical Thinking
  • Data Analysis
  • Data Privacy Standards
  • Data Quality
  • Data Reporting
  • Data Savvy
  • Data Science
  • Data Visualization
  • Digital Fluency
  • Econometric Models
  • Organizing
  • Process Improvements
  • Strategic Thinking
  • Technical Credibility
  • Workflow Analysis

What the JD emphasized

  • 5+ years of industry experience delivering on data science projects using ML predictive modeling, multi-objective optimization, natural language processing, and GenAI.
  • Hands-on experience with multi-modal ML predictive modeling and stochastic simulations for time-series forecasting.
  • Experience building multi-objective optimization engines to navigate complex trade-offs using evolutionary algorithms, reinforcement learning, or mixed-integer linear programming.
  • Experience with GenAI and clinical LLMs for document parsing and clinical concept disambiguation and harmonization.

Other signals

  • Leverage operational, RWD, and cost data to build ML predictive models and optimization engines
  • Adapt large language models (LLMs) for tailored information extraction
  • Generate actionable insights, enabling early intervention and risk management