Numerical Program Analysis for Scientific Computing Postdoctoral Researcher
Ora
Summary
We have an opening for a Postdoctoral Researcher to advance compiler-based techniques for detecting and understanding floating-point errors in scientific applications. You will explore extending existing LLVM-based instrumentation and analysis frameworks while also investigating new instrumentation architectures and independently developed tools. You will research error-tracking techniques and evaluate them on large-scale scientific applications running on high-performance computing systems. Research topics may include identifying numerical instabilities and supporting mixed-precision optimization. A key component of this work will be using and developing AI-agent workflows to accelerate research and software development. You will investigate how AI agents can work effectively with compiler and runtime analyses while ensuring that their results remain reproducible, explainable, and rigorously validated. You will collaborate with programming-languages researchers, HPC software developers, numerical experts, and Laboratory application teams. This position is in the Center for Applied Scientific Computing (CASC) within the Computing Directorate.
Job description
In this role you will
- Research program-analysis techniques for detecting, explaining, and mitigating floating-point errors in scientific applications.
- Explore extensions to existing LLVM-based floating-point instrumentation frameworks.
- Design new LLVM-based instrumentation and analysis tools when alternative approaches would improve functionality, scalability, or usability.
- Develop and use AI agents to accelerate compiler-tool development, code analysis, test generation, experimentation, and numerical-error diagnosis.
- Develop methods to validate AI-generated code, analyses, and recommendations for correctness and reproducibility.
- Investigate applications of floating-point error-tracking analysis, including mixed-precision optimization and detection of numerically unstable code.
- Collaborate with Laboratory scientists to deploy and evaluate the research on real-world applications.
- Pursue independent (but complementary) research interests and being able to interact with team members internally and externally.
- Publish research results in external peer-reviewed scientific conferences and journals.
- Perform other duties as assigned.
Qualifications
- Ph.D. in Computer Science, Computer Engineering, Computational Science, or a related field.
- Demonstrated experience developing compiler transformations or program-analysis tools using LLVM or comparable compiler infrastructure.
- Demonstrated experience with floating-point analysis, numerical accuracy or stability, mixed-precision computation, or closely related analysis of numerical scientific software.
- Experience building, running, or analyzing scientific applications in an HPC environment, including familiarity with parallel programming models such as MPI, OpenMP, CUDA, or similar technologies.
- Experience developing systems or scientific software in C or C++ in a Linux environment.
- Ability to conduct high-quality research and implement sophisticated techniques for experimental evaluation.
- Demonstrated research productivity through peer-reviewed publications, technical reports, presentations, and/or open-source software.
- Proficient written and verbal communication skills necessary to collaborate with multidisciplinary research teams, author scientific papers, and deliver technical presentations.
- Ability to independently identify and solve complex research and software-development problems.
Qualifications We Desire
- Knowledge of floating-point arithmetic.
- Experience with dynamic analysis, shadow execution, sanitizers, debugging tools, or performance-analysis tools.
- Experience with mixed-precision algorithms, precision tuning, or automated optimization of numerical programs.
- Experience with formal methods, abstract interpretation, symbolic execution, or related software-verification techniques.
- Experience validating AI-generated code and analyses through automated testing, runtime evidence, and reproducible evaluation.
Position Information
This is a Postdoctoral appointment with the possibility of extension to a maximum of three years. The postdoctoral appointment is a temporary, mentored training position intended for individuals in the early stages of their postdoctoral research careers. Consistent with the structure of the program, applicants must have received their Ph.D. (or equivalent doctorate) within five years of the anticipated start date.
Requirements
- Ph.D. in Computer Science, Computer Engineering, Computational Science, or a related field
- Demonstrated experience developing compiler transformations or program-analysis tools using LLVM or comparable compiler infrastructure
- Demonstrated experience with floating-point analysis, numerical accuracy or stability, mixed-precision computation, or closely related analysis of numerical scientific software
- Experience building, running, or analyzing scientific applications in an HPC environment, including familiarity with parallel programming models such as MPI, OpenMP, CUDA, or similar technologies
- Experience developing systems or scientific software in C or C++ in a Linux environment
- Ability to conduct high-quality research and implement sophisticated techniques for experimental evaluation
- Demonstrated research productivity through peer-reviewed publications, technical reports, presentations, and/or open-source software
- Proficient written and verbal communication skills necessary to collaborate with multidisciplinary research teams, author scientific papers, and deliver technical presentations
- Ability to independently identify and solve complex research and software-development problems
- Knowledge of floating-point arithmetic
- Experience with dynamic analysis, shadow execution, sanitizers, debugging tools, or performance-analysis tools
- Experience with mixed-precision algorithms, precision tuning, or automated optimization of numerical programs
- Experience with formal methods, abstract interpretation, symbolic execution, or related software-verification techniques
- Experience validating AI-generated code and analyses through automated testing, runtime evidence, and reproducible evaluation
Benefits
- 401(k)
- Relocation Assistance
- Education Reimbursement Program
- Flexible Schedules
- Flexible Benefits Package
Company overview
Ora is the world's leading full-service ophthalmic drug and device development firm, operating as a specialized contract research organization (CRO) focused exclusively on ophthalmology. Founded in 1974 and headquartered in Andover, Massachusetts, the company has over 45 years of experience in advancing eye care innovations globally. Ora provides comprehensive research and development services spanning preclinical studies, clinical trials, regulatory strategy, and product development for both pharmaceutical and medical device companies. With operations across the United States, United Kingdom, Australia, and Asia, the company has helped clients achieve more than 85 product approvals to date. Ora's expertise covers a wide range of ophthalmic conditions including allergy, dry eye, retina diseases, glaucoma, and anti-infective therapies. The company employs over 500 ophthalmology specialists who utilize proprietary clinical models, integrated data solutions, and global regulatory strategies to accelerate product development from concept to market. Ora serves a diverse client base ranging from startups to global pharmaceutical and device companies, positioning itself as the trusted partner for ophthalmic innovation worldwide.