Postdoctoral Fellow – AI-Driven Multi-Omics Integration for Predictive Toxicology Part-time NEW

Postdoctoral Fellow In Ai-Driven Multi-Omics Integration For Predictive Toxicology

Pfizer's Drug Safety Research and Development (DSRD) team is actively seeking a postdoctoral fellow in AI-driven multi-omics integration for predictive toxicology—an opportunity to push the boundaries of AI, biology, and drug safety innovation.

The postdoctoral fellow will develop and apply foundation model (FM) and machine learning approaches to integrate multi-omics data—including transcriptomics and proteomics—generated from preclinical in vitro and in vivo safety studies. The fellow will benchmark biological foundation models (e.g., scGPT, GeneFormer) alongside linear and classical ML baselines against curated cross-species toxicology datasets, build end-to-end AI pipelines that connect early omics readouts to downstream pathology, clinical chemistry, and other endpoints to uncover subtle biological signals predictive of human toxicity.

By leveraging cutting-edge AI methods, the project aims to identify novel molecular biomarkers and early indicators of drug-induced safety liabilities, enabling cross-species prediction of human-relevant safety risks from preclinical data. This research will directly support predictive toxicology and translational safety decisions in drug development, helping inform go/no-go and de-risking strategies.

Progress in this area will be driven by strong scientific contributions—peer-reviewed publications, conference presentations, and potentially open-source analytical tools—to ensure the impact of this work both within Pfizer and in the broader scientific community. The ideal candidate will have a solid background in computational biology, bioinformatics, or a related field, proficiency in AI/ML techniques, and a passion for applying cutting-edge models to real-world biomedical data in order to advance drug safety science.

Role responsibilities:

• Benchmark foundation and linear models against a curated cross-species omics dataset library spanning decision-relevant toxicity endpoints (liver, cardiac, hematopoietic), defining performance criteria that are meaningful for safety go/no-go decisions.
• Develop and validate AI-driven integration pipelines that combine multi-omics data from early toxicology studies with historical endpoints—pathology scores, clinical chemistry, and PK data—using foundation models and interpretable ML approaches.
• Perform retrospective compound analyses to quantify where early omics-based model outputs could have anticipated findings from GLP or Phase 1 studies and prospectively integrate targeted in vitro datasets to measure the incremental predictive value of each data modality.
• Implement scalable Python and/or R workflows for data ingestion, model training, evaluation, and visualization, including APIs or interactive applications to support internal stakeholder adoption
• Collaborate with toxicologists, pathologists, data scientists, and external partners to integrate in silico, in vitro, and in vivo results into translational safety frameworks.
• Communicate findings through internal reports, peer-reviewed publications, conference presentations, and open-source software releases to influence both internal safety processes and the broader field of computational toxicology.

Basic qualifications:

• Ph.D. in computational biology, bioinformatics, computational toxicology, systems biology, or a related scientific field.
• Experience: 0–2 years postdoctoral or post-PhD research experience (candidate must have completed doctorate within the past 2 years). Willingness to make a minimum 2-year commitment to the fellowship (fixed-term role). Two letters of recommendation will be required for final interviews.
• Scientific contributions: Demonstrated research productivity with at least one first-author publication published or submitted in a peer-reviewed journal.
• Technical skills: Proficiency in programming and data analysis using Python and/or R. Strong statistical and machine learning skills for analyzing complex biological datasets.
• Domain knowledge: Familiarity with high-dimensional biological data (such as transcriptomics, genomics, proteomics) and a basic understanding of molecular biology or toxicology to contextualize data-driven findings.
• Advanced AI/ML expertise: Hands-on experience with modern deep learning frameworks (e.g. PyTorch, scikit-learn) and large-scale machine learning models. Exposure to representation learning, transformer-based architectures, or self-supervised learning on biological or biomedical data is a strong plus.
• Multi-omics & bioinformatics: Experience with single-cell or bulk RNA-seq analysis pipelines (e.g. Scanpy, Seurat, DESeq2).

Preferred qualifications:

• Translational insight: Knowledge of toxicology, pharmacology, or biomarker discovery—for example, understanding common preclinical safety study endpoints or translational research—is an advantage. Ability to interpret and validate model results in a biological/toxicological context will be valuable.
• Collaboration & communication: History of interdisciplinary collaboration and strong communication skills. Experience working in cross-functional research teams or with external collaborators (academia or industry) on complex data projects is a plus.

Physical/mental requirements:

Ability to perform complex data analysis and quantitative modeling tasks. Must be able to concentrate on computational work (e.g. coding, data interpretation) for extended periods and perform precise, detail-oriented mathematical calculations as required.

Additional information:

Relocation support available work location assignment: hybrid. Last day to apply may 24, 2026.