Sr. Machine Learning Researcher, Domain-Aware Modeling & Scientific Machine Learning

Bayer•Tulsa, OK

3d•$120,000 - $170,000•Remote

About The Position

We are seeking a Sr. Machine Learning Researcher with strong expertise in the mathematical foundations of machine learning and scientific computing to develop next-generation domain-aware models for agriculture. This role sits at the intersection of applied mathematics, domain-aware modeling, and deep learning, with the goal of building models that respect and encode the underlying structure of biological and environmental systems. You will design principled, interpretable, and generalizable AI architectures that integrate scientific knowledge from genetics to crop physiology to environmental dynamics- into data-driven frameworks. Your work will directly enable transformative applications in genomic selection and genome editing target identification, accelerating the development of improved crop varieties worldwide.

Requirements

PhD in one of the following or closely related fields: Machine Learning / Deep Learning, Applied Mathematics, Computational Science & Engineering, Physics, Chemical, Mechanical, or Biomedical Engineering, Computer Science (with scientific computing or numerical methods focus), Statistics / Probabilistic Modeling, Another related quantitative discipline with demonstrated depth in mathematical modeling.
Demonstrated research output (publications, thesis work, or applied projects) in scientific machine learning, numerical methods for differential equations, or data-driven modeling of physical/biological systems.
Proficiency in modern deep learning frameworks (PyTorch, JAX, or TensorFlow) and scientific computing libraries.
Experience formulating and solving problems involving high-dimensional, structured, or multi-modal data.
Strong communication skills and willingness to collaborate across disciplines.

Nice To Haves

5+ years post-PhD relevant experience
Demonstrated experience with one or more of the following domain-aware modeling paradigms: Physics-Informed Neural Networks (PINNs), Biology-Informed Neural Networks (BINNs) / Visible Neural Networks (VNNs), Neural Ordinary/Partial Differential Equations (Neural ODEs/PDEs), Operator learning methods (e.g., DeepONet, Fourier Neural Operator), Hybrid mechanistic–data-driven models
Experience with Bayesian inference, Gaussian processes, hierarchical models, or probabilistic programming.
Familiarity with nonlinear dynamics, dynamical systems theory, or systems biology modeling.
Background in surrogate modeling, model reduction, or multi-fidelity methods.
Exposure to genomics data structures (e.g., variant matrices, linkage disequilibrium, population genetics) or quantitative genetics (e.g., genomic BLUP, marker-effect models) - not required, but valued.
Experience deploying ML models into production environments (MLOps, containerization, cloud-based HPC).
Experience collaborating in interdisciplinary research teams spanning experimental and computational scientists.
Familiarity with ensemble methods, gradient-boosted models, kernel methods, or classical statistical learning as complementary tools.

Responsibilities

Design, build, and validate domain-aware machine learning models (e.g., biology-informed, and hybrid mechanistic-statistical architectures) that incorporate prior scientific knowledge into learning algorithms for agricultural and genomic applications.
Develop novel architectures and loss functions that embed biological constraints, conservation laws, symmetry properties, or known functional relationships into neural network training, ensuring physically and biologically consistent predictions.
Architect models that leverage high-dimensional genomic, phenomic, and environmental data to predict complex trait outcomes, identify causal genetic variants, and prioritize genome editing targets with quantified uncertainty.
Implement rigorous uncertainty quantification frameworks (Bayesian deep learning, ensemble methods, probabilistic surrogate models) to provide decision-makers with calibrated confidence estimates on model predictions.
Partner with geneticists, plant biologists, agronomists, environmental scientists, and software engineers to translate domain expertise into model architecture decisions and validate model outputs against biological ground truth.
Work with engineering and IT teams to transition research prototypes into production-grade models integrated within breeding and discovery pipelines, ensuring reproducibility, scalability, and maintainability.
Contribute to publications in leading venues, participate in the internal scientific community, and stay at the frontier of scientific machine learning methodology.
Prepare comprehensive technical documentation, present findings to both technical and non-technical stakeholders, and build organizational trust in AI-driven decision-making.