Principal Scientist, Predictive Sciences

GSK•Upper Providence, PA

About The Position

The Principal Scientist will work in the Predictive Sciences group within GSK’s CMC Digital & Knowledge Management department, at the intersection of late-phase biologics CMC process development and digital tools, predictive solutions, and artificial intelligence. The role advances a predict-first, digitally augmented, model-informed approach to biologics development and supports upstream and downstream processes across fed-batch, intensified fed-batch, and continuous perfusion platforms. This individual contributor role calls for strong modeling expertise alongside the ability to deploy models into tools that scientists can use independently. The Principal Scientist will take mechanistic, hybrid, and data-driven models from prototype to production, using machine-learning workflows and agentic AI systems to inform development decisions. The work also supports technology transfer, scale-up, and commercialization, and requires leading digital transformation and change across the business. The role reports to the Team Lead for LM Drug Substance Development modelling. It partners closely with process development, analytical development, manufacturing science, and broader GSK digital science and data communities. The Principal Scientist is expected to exercise high scientific autonomy and influence digital CMC through technical contributions and trusted scientific partnerships.

Requirements

PhD in data science, computational science, AI/ML, chemical engineering, biochemical engineering, computational biology, or a closely related field; alternatively, a Master's degree with at least 2 years of industry experience.
Experience building predictive bioprocess models, including mechanistic, hybrid, data-driven, or chemometric models.
Experience with Python and scientific computing tools such as NumPy, SciPy, or PyTorch, or equivalent tools including MATLAB, gPROMS, or SIMCA.
Experience with CMC development principles, including process development, analytical characterization, scale-up, and biologics regulatory frameworks such as ICH Q8/Q9/Q10/Q11.
Experience applying machine learning to small-to-medium scientific datasets, with understanding of overfitting, uncertainty quantification, and model interpretability.
Experience taking models into production, including deployment, versioning, monitoring, and MLOps practices such as testing, packaging, and CI/CD, with the ability to deliver models for independent use by non-modeling end users.
Experience building or working with agentic AI frameworks in a scientific or industrial context.

Nice To Haves

Familiarity with Process Analytical Technology (PAT) and real-time process control.
Experience with Design of Experiments and multivariate data analysis in a biopharmaceutical setting.
Exposure to regulatory filings such as IND, BLA, or NDA and the role of models in CMC submissions.
Knowledge of digital twin platforms or process simulation tools such as gPROMS, MATLAB, or Aspen.

Responsibilities

Build and deploy mechanistic, hybrid, and data-driven process models across upstream cell culture, chromatography, filtration, and formulation, including digital twins and PAT chemometric models.
Use engineering, biological, chemical, or materials science knowledge to propose mechanisms for bioprocess behavior and develop or adapt mathematical models.
Drive model-informed process development and characterization, using MBDoE and model-based risk assessments to define critical process parameters, design space, in-silico experiments, platform fit, monitoring, and control strategies.
Translate mathematical models into working computational tools using platforms such as gPROMS, SIMCA, and JMP, or custom code such as Python and MATLAB; validate models against internal compliance and regulatory requirements.
Champion adoption of digitally redesigned CMC workflows across Drug Development and Supply, enabling routine model use by function scientists through documentation and operating procedures.
Build and sustain shared modeling infrastructure, tools, and practices that keep models versioned, monitored, reliable, and reusable; embed data integrity, quality, safety, policy, and intellectual-property protections.
Advance capabilities through publications, conferences, academic collaborations, pre-competitive alliances, and external expertise.
Present results to CMC teams and senior leadership, and collaborate across R&D, pilot plants, and the commercial supply chain.
Design and build agentic AI systems that analyze historical CMC data, propose experimental designs and strategies, invoke models, and provide actionable recommendations.
Integrate large language models and AI agents into CMC workflows, including literature mining, predictive intelligence, process insights, and development acceleration.
Develop AI-driven risk assessment and criticality-ranking tools for process characterization and control strategy definition.
Work with laboratory informatics, data engineering, and technology platform teams to link digital solutions with electronic lab notebooks, process historians, and analytical instruments.
Champion FAIR data principles and structured data capture across CMC development teams; contribute to shared modeling infrastructure, reusable components, and validated computational workflows.
Partner with CMC teams to translate biological and process complexity into computable problems and model outputs into practical development decisions.
Present modeling results and AI-derived insights to senior stakeholders, regulatory teams, and cross-functional partners.
Author or co-author technical reports, model qualification packages, and regulatory submission sections as needed.