Postdoctoral Researcher

University Health Network • Toronto, ON

6d • CA$54,902 - CA$93,333 • Onsite

About The Position

We are seeking a postdoctoral researcher with strong expertise in AI/ML to join a major interdisciplinary initiative focused on developing foundation models and secure computational infrastructure for translational cancer research. The first major objective is to build a general-purpose drug foundation model capable of transfer learning across diverse prediction tasks, including mechanism of action classification, clinical drug response prediction in tumour subtypes, ADMET and toxicity profiling, combinatorial drug synergy, and drug repurposing. The goal is to move beyond task-specific architectures and datasets toward flexible models that can generalize across therapeutic contexts. The second major objective is to develop and apply secure, scalable, and privacy-preserving computational infrastructure to support biomarker discovery across diverse treatment modalities. This will include building agentic AI approaches to harmonize clinical, genomic, and transcriptomic data across public and private cohorts, while assessing the predictive value of DNA and RNA signatures in federated settings where sensitive data remain under local governance. Together, these efforts aim to advance AI-driven drug discovery, biomarker development, and clinical translation by integrating modern machine learning, multimodal biomedical data, and robust distributed analysis frameworks. The successful candidate will work in the Haibe-Kains Lab at the Princess Margaret Cancer Centre, University Health Network.

Requirements

• Awarded a PhD within the previous 5 years, or an MD or DDS within the previous 10 years in a relevant quantitative or biomedical discipline, including but not limited to: Computational Biology, Bioinformatics, Systems Biology, or Quantitative Genomics; Machine Learning, Artificial Intelligence, Computer Science, or Data Science; Biostatistics, Biomedical Engineering, or related fields
• Demonstrated experience developing or applying deep learning methods to molecular, biological, clinical, or multi-omics data.
• Expertise in one or more modern AI/ML approaches relevant to foundation models or representation learning, such as graph neural networks, transformers, generative models, self-supervised learning, few-shot or zero-shot learning, or transfer learning.
• Strong programming skills in Python and/or R, with practical experience using modern machine learning frameworks and tooling such as PyTorch, Hugging Face, PyTorch Geometric, Deep Graph Library, scikit-learn, or equivalent platforms.
• Experience working with large-scale biomedical datasets, such as molecular graphs, transcriptomic perturbation profiles, multi-omics data, clinical genomics, electronic health record-derived data, or treatment-response datasets.
• Proficiency with reproducible workflow management systems such as Snakemake, Nextflow, CWL, or equivalent pipeline frameworks.
• Familiarity with cloud or high-performance computing environments, such as GCP, AWS, SLURM-based clusters, or equivalent infrastructure.
• Strong publication record, commensurate with career stage, in computational biology, AI/ML, bioinformatics, biostatistics, biomedical data science, or related fields.
• Excellent communication skills and ability to work collaboratively in interdisciplinary teams spanning computational biology, machine learning, software engineering, oncology, and clinical research.

Nice To Haves

• Understanding of data harmonization, privacy-preserving analysis, federated learning, secure distributed computing, or clinical data governance is highly desirable.
• Experience with clinical data harmonization (e.g., OMOP CDM, HL7 FHIR) is preferred.
• Experience designing scalable bioinformatics pipelines for large-scale genomic or transcriptomic datasets, preferred.
• Understanding of regulatory and data governance requirements in clinical research settings, preferred.
• Prior collaborative work across computational and clinical or wet-lab research teams, preferred.

Responsibilities

• Design and implement multimodal drug foundation models that integrate molecular graph representations, bulk transcriptomic perturbation signatures, and multi-omics cell-state representations.
• Develop transfer learning strategies to support diverse drug prediction tasks, including mechanism of action classification, clinical drug response prediction, ADMET and toxicity profiling, combinatorial drug synergy, and drug repurposing.
• Build flexible AI/ML workflows that reduce reliance on task-specific architectures and enable generalization across therapeutic contexts, tumour subtypes, and treatment modalities.
• Architect and deploy secure, scalable, and privacy-preserving computational infrastructure for biomarker discovery and translational cancer research.
• Develop agentic AI-enabled frameworks to support the harmonization, annotation, quality control, and integration of clinical, genomic, and transcriptomic data across public cohorts and private institutional datasets.
• Implement distributed and federated analysis pipelines in which each contributing dataset can be analysed separately, enabling multi-cohort biomarker assessment without raw data centralization.
• Develop systematic workflows to evaluate published and user-specified DNA and RNA signatures, including immune, stromal, mutation-based, pathway-level, and treatment-response signatures.
• Assess the predictive value of molecular signatures across cancer types and treatment modalities, including chemotherapy, targeted therapy, immunotherapy, and emerging therapeutic approaches.
• Integrate biomarker discovery and immunotherapy inference pipelines with clinical data warehouses to support translational studies in collaboration with clinical, industry, and computational partners.
• Contribute to responsible data sharing frameworks, data governance processes, IRB/ethics protocols, and regulatory documentation as required.
• Collaborate closely with computational biologists, software developers, clinicians, and wet-lab scientists to build, validate, and translate predictive models and biomarker discovery workflows.

Benefits

• Competitive offer packages
• Government organization and a member of the Healthcare of Ontario Pension Plan (HOOPP)
• Close access to Transit and UHN shuttle service
• A flexible work environment
• Opportunities for development and promotions within a large organization
• Additional perks (multiple corporate discounts including: travel, restaurants, parking, phone plans, auto insurance discounts, on-site gyms, etc.)

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: Ph.D. or professional degree