CISL Postdoctoral Fellow I: Large Language Models for Code Modernization

Visa Sponsored Job:

Relocation Assistance Eligible:

Job Location:

Position Type & Term:

$80,480

*Final salary and rates are based on education, experience, and skills relevant to the role.*

Salary Annual Rate $80,480

*Final salary and rates are based on education, experience, skills relevant to the role.*

Application Deadline: This position will be posted until 11:59:59pm, Thursday, May 15, 2025. Applications will not be accepted past this date.

Required application materials (preferably in PDF Format):

• Resume

• Cover Letter - Please address how your skills and experience meet the needs of this position (for more information, please refer to the Key Responsibilities and Knowledge, Skills, and Abilities sections of this job posting).

  • Our hiring team wants to learn more about your skills/experiences regarding very specific topics related to the critical skills needed to be successful in this position. Your responses to these topics will be read and weighed equally to your resume, so please ensure you're providing detailed, yet informative, responses based on your previous work experiences.

  • Please select 2 of the following topics to expand upon in your cover letter, and keep your responses to 1-2 paragraphs.

    • Experience with open-source LLMs, such as LLaMA, Mistral, Falcon, or similar frameworks (Hugging Face, LangChain, etc.), especially in the context of fine-tuning, evaluation, or deployment.

    • Hands-on work with production-level code, including familiarity with software engineering practices such as version control, testing, and refactoring.

    • Research or project work in code comprehension or automated code generation, including examples of tasks like bug fixing, transformation, or modernization using AI tools.

    • Approach to explainable AI, especially methods used to interpret model outputs or transformation decisions.

    • Contributions to open-source projects or publications, particularly those involving LLMs, AI for software engineering, or interpretable ML.

Relocation Assistance Eligible: Partial relocation available (must live beyond a 50-mile radius of a UCAR operating site).

Position Type & Term:

This is a full-time, 2 year term position with the possibility of extension.

Background checks are conducted for candidates selected for hire. Learn more.

Work Location Expectations: This position is open to candidates seeking in-person and/or hybrid (combination of 3 days in-person and 2 days of remote) opportunities.

Here is a summary of what one would expect to be generally responsible for in this role.

Key Responsibilities

Advanced AI/ML Research & Scientific Communication

• Develop analytical, generative, and transformative AI models for software development and code modernization.

• Validate accuracy, performance, and reliability of LLM-based code transformation and the quality of uncertainty estimation methods.

• Analyze models to understand effectiveness of code modernization with interpretable explainable AI techniques.

• Present results at project meetings and conferences; publish results in the peer-reviewed scientific literature, and review scientific manuscripts.

Collaborative Leadership

• Contribute to team meetings and discussions, general coordination with NSF NCAR team.

• In collaboration with NSF NCAR team members, may assist with mentoring students, interns, and visitors.

Successful candidates will ensure their application materials speak to the following criteria:

Education & Experience

• PhD in computer science, atmospheric science, physics, engineering, or other relevant science.

Knowledge, Skills, and Abilities

• Demonstrated knowledge of training and fine-tuning large language models (with at least one major deep learning framework (e.g., PyTorch, Keras, TensorFlow, JAX)).

• Demonstrated knowledge of Python for software development and code analysis.

• Ability to process and analyze large codebases and software repositories.

• Skilled in performing computing tasks on Linux computing clusters.

• Excellent oral and written communication skills with demonstrated success publishing in peer reviewed journals and/or computer science conferences.

• Ability to work both independently and as part of a larger team.

• Skill in planning and managing project goals, objectives, and priorities to meet goals in a timely manner.

• Demonstrated creative scientific insight.

Desired, but not required:

• Knowledge of software engineering principles and code modernization techniques.

• Knowledge of uncertainty quantification methods in AI-based code generation and how to validate them.

• Knowledge of explainable AI methods for code transformation and how to implement them.

• Skills in distributed training of LLMs across multiple GPUs.

UCAR affirms its commitment to employees through competitive benefits. In addition to medical, dental, vision, retirement, and life insurance, UCAR offers a variety of programs focused on work-life balance and professional, and personal development. These include:

• Tuition Assistance, time off allowance to attend classes, and other professional development opportunities

• UCAR contributes 10% of your eligible pay into your retirement account; 100% fully vested on day one

• Starting minimum accrual of 20 days of personal time off each year (prorated for less than full-time positions)

• 10 paid holidays

• 10 days of sick leave each year

• 12 weeks of paid parental leave

• Short-term medical leave paid at 100% of your regular salary

• EcoPass for local Colorado residents to use the Denver and Boulder-area transit system at no cost

Applicants are not required to provide age or age-related information and may redact information related to age, date of birth, or dates of attendance at or graduation from an educational institution from any submissions during the initial application process.

At UCAR|NCAR|UCP, you will work alongside a dedicated team of professionals conducting critical research and community outreach to solve complex Earth system science problems including climate change, air pollution, extreme weather, floods, drought, wildfires, and space weather, all with the goal of improving human life and reducing economic loss. Each of us, from scientists to the professionals who support their work, serves the public and a collaborative community of scientists in our mission to understand the complex processes that make up the Earth system, from the ocean floor to the Sun's core.

Flexible Work

At UCAR, we are committed to supporting our mission by giving staff the flexibility to find the schedule and location that works best to maintain their own work-life circumstances and reach their full potential as professionals. Many positions within our organization are eligible for fully on-site, hybrid (three days per week) and/or flexible work hours.

Equal Opportunity Employer

UCAR is committed to providing equal opportunity for all employees and applicants for employment and does not discriminate on the basis of race, age, creed, color, religion, national origin or ancestry, sex, gender, disability, veteran status, genetic information, sexual orientation, gender identity or expression, or pregnancy.Whatever your intersection of identities, you are welcome at UCAR.

Export Control

All positions are required to comply with U.S. export compliance regulations work location requirements regarding access to facilities and research systems.

Visa Wait Times

Please consider the length of visa procurement when applying for this posting, understanding that you will not be able to begin employment until you are able to get a visa and enter the U.S.

Work Location

UCAR requires ALL positions to be performed within the U.S., excluding U.S. Territories.

AI Software

ChatGPT and similar AI software are powerful tools that are changing theway society receives, processes, and leverages information promptly. While we acknowledge its benefits and do not restrict leveraging it with job applications, we highly encourage a majority of the applicant material to be original work.