Postdoctoral Fellow, Structural Biology

State University of New York at Buffalo; New York (NY); On-site; USD 70,000 - 85,000; Full time

Job summary

A prestigious university is seeking a highly skilled Postdoctoral Fellow to lead the computational development of a novel generative AI framework for structural biology. This role focuses on integrating X-ray scattering data to model protein ensembles, utilizing advanced GPU workflows. The position requires a completed doctoral degree in a related field and advanced proficiency in PyTorch and distributed training techniques. A competitive salary range of $70,000 - $85,000 is offered for this full-time position based at the university's downtown campus.

Qualifications

• Completed doctoral degree requirements by start date.
• Strong understanding of model stability and gradient accumulation.
• Familiarity with geometric deep learning or diffusion architectures.

Responsibilities

• Lead technical evolution of project for protein ensemble modeling.
• Implement next phase from rigid-body models to sophisticated systems.
• Resolve hardware-specific performance challenges across GPU environments.
• Design new loss functions and data integration from experimental data.

Skills

Expert-level proficiency in PyTorch

Expert-level proficiency in JAX

Experience with Distributed Training

Managing large-scale GPU workloads

Understanding of mixed-precision training

Education

Doctoral degree in Computer Science, Data Science, or related field

Tools

CUDA for hardware-specific performance

Docker

Singularity

Job description

The Grant Lab at the University at Buffalo is seeking a highly skilled Postdoctoral Fellow to lead the computational development of a novel generative AI framework for structural biology. This project sits at the intersection of X‑ray scattering and deep learning, aimed at integrating experimental data to predict protein ensemble structures. As an Empire AI‑funded fellow, you will have early access to the Empire AI clusters, utilizing state‑of‑the‑art GPU architectures to push the boundaries of structural biology.

This position is a prestigious Empire AI Fellowship at the University at Buffalo, designed for a “CS‑first” researcher to drive the technical evolution and large‑scale implementation of a new platform for protein structure prediction. Working at the intersection of generative AI and biophysics, the Fellow will focus on expanding the current framework to model dynamic protein ensembles. You will have access to the Empire AI Alpha and Beta clusters, utilizing hundreds of state‑of‑the‑art GPUs (including H100 and GB200 nodes) for scaling generative models for structural biology.

Responsibilities

• Technical Expansion: Implement the next phase of the project to transition from rigid‑body models to sophisticated systems for protein ensemble modeling.
• Computational Optimization: Resolve hardware‑specific performance and numerical precision challenges across diverse GPU environments.
• Architecture Design: Lead the design of new loss functions, model architectures, and synthetic datasets that integrate experimental X‑ray scattering data.
• System Stability: Ensure numerical reproducibility and stability in large‑scale distributed training workloads.

As an Equal Opportunity / Affluent Action employer, the Research Foundation will not discriminate in its employment practices due to an applicant’s race, color, religion, sex, sexual orientation, gender identity, national origin and veteran or disability status.

Minimum Qualifications

• Doctoral degree or equivalent in Computer Science, Data Science, Computational Physics, or a related field with a focus on Deep Learning.
• All degree requirements, including dissertation, must be completed by the start date.
• Expert‑level proficiency in PyTorch and/or JAX.
• Demonstrated experience with Distributed Training (e.g., DeepSpeed, FSDP) and managing large‑scale GPU workloads.
• Strong understanding of low‑level model stability, including mixed‑precision training (BF16/FP8) and gradient accumulation.

Preferred Qualifications

• At least two years of experience beyond the PhD in a research or engineering environment focused on large‑scale AI.
• Experience with geometric deep learning, diffusion architectures, or related frameworks (e.g., OpenFold, AlphaFold2/3).
• Familiarity with Docker/Singularity for reproducible HPC environments.
• Experience with CUDA‑level optimization or debugging hardware‑specific performance differences.
• Basic knowledge of protein structure, folding, or biophysics.

Salary Range

Salary Range: $70,000 - $85,000

Work Hours

Work Hours: 37.5 hours per week

Campus

Campus: Downtown Campus