At Fujitsu, we are driven by our purpose to make the world more sustainable by building trust in society through innovation.

We have been a pioneer in technology and innovation for over 80 years, and we are committed to using our expertise to help businesses and organizations transform for the digital age. We believe that digital transformation is essential to creating a more sustainable future. That's why we are working with our customers to develop solutions that can help them reduce their environmental impact, improve their efficiency, and create a more equitable society.

We are committed to contributing to the United Nations Sustainable Development Goals (SDGs). These goals are a blueprint for a better future for all, and we believe that technology can play a vital role in achieving them.

If you share our passion for making a meaningful impact on the world, we invite you to join our global family of 130,000 employees spanning more than 50 countries. We are a diverse workforce, and we offer a wide range of opportunities for you to grow and develop your career.

Together, we can create a more sustainable future for all.

FUJITSU RESEARCH OF AMERICA, INC. (FRA) was established in 1993 as a wholly owned subsidiary of Fujitsu Laboratories Ltd., which is a wholly owned subsidiary of Fujitsu Ltd., focusing on contributing to Fujitsu's technology value chain.  FRA conducts advanced research in Artificial Intelligence, Data Science, System Components and Platform Innovations, Network and Trusted Systems Innovations, and Internet Service Innovations. 

We are seeking a highly motivated PhD intern to join our team working at LLM-based coding agents. This internship will focus on building intelligent agents and fine-tuning LLMs that assist in end-to-end coding workflows and on benchmark evaluations.

You will collaborate closely with a team of AI researchers and engineers, contribute to system design and experimentation, and help shape the future of automated software development.

Job Description 

The candidate will work on:

• Build and evaluate LLM-based coding agents.
• Improve agent performance with both training and test-time methods.
• Deliver reproducible results, with the goal of submitting a publication to a top AI conference.

Qualifications 

• Currently pursuing a PhD in Computer Science, Machine Learning, Artificial Intelligence, or a related field.
• Research experience with LLMs and/or agentic tool use.
• Strong Python and Java; comfortable with git, debugging, unit tests, and Linux/CLI workflows.
• Experience running reproducible experiments (baselines, ablations, clean reporting).
• Publication record in top AI/ML venues

Preferred Qualifications (Plus)

•  Hands-on experience with coding agent benchmarks like SWE-bench, Terminal-Bench, CodeClash, etc.
• Experience fine-tuning LLMs for code (e.g., SFT, DPO/RLHF-style methods, RL) and building training data/trajectories.
• Experience with verification at inference time (unit tests, static analyzers/linters/type-checkers, learned verifiers, search/reranking).
• Experience with Docker/containers, CI-style evaluation, and scaling experiments.

Fujitsu aligns salaries to the specific geographic location in which the work is primarily performed. It is not typical for an individual to be hired at or near the top of the range for their role and compensation decisions are dependent on the circumstances of each situation. The pay range for this role takes into account the wide range of factors that are considered in making compensation decisions including but not limited to: specific skills, qualifications, experience, and comparison to other employees already in this role. The pay range for this position is estimated at $35/hr. to $50/hr. USD.