Research Software Engineer, UD Institute for Data Science

University of Delaware•Newark, NJ

About The Position

The University of Delaware (UD) invites applications for a Research Software Engineer (RSE) to join the NSF AI Institute for Human-AI Cooperation (HAIC). The HAIC Institute is a 5-year, multi-institutional research initiative supported by the National Science Foundation. HAIC brings together researchers and partners to develop trustworthy AI that strengthens human expertise, shapes the future of rehabilitation and health care, and cultivates the next generation of leaders in AI. The RSE will work with HAIC researchers from UD and partner institutes to build, validate, and maintain robust and scalable open-source software tools and cyberinfrastructure that support institute-wide research in artificial intelligence, biophysical digital twins, multimodal data processing, personalized rehabilitation, and human-AI symbiosis. The RSE will leverage high-performance computing (HPC) systems and work extensively with research computing professionals and NVIDIA systems engineers. Software needs include supporting reproducible research, interfacing with real-time systems and databases, and translation of AI models for a variety of deployment settings. The RSE will join a growing cohort of RSE at UD, with coordinated professional development opportunities. This position is supported by the U.S. National Science Foundation.

Requirements

Bachelor’s degree in computer science, electrical and computer engineering, data science, statistics, biomedical engineering, mechanical engineering, operations research, robotics, computational neuroscience, applied mathematics, physics, or a closely-related discipline and five years of related work experience.
Experience with the Linux operating system and utilizing the command line interface.
Demonstrated ability and experience with scripting and programming languages (Bash, Python).
Hands-on experience with shared high-performance computing environments and workload scheduling tools (e.g., SLURM).
Extensive experience with version control and git, with contributions to open source repositories.
Ability to work effectively with researchers (both computational and non-computational) within multi-institutional teams to define, clarify, and execute requirements and goals.
Excellent oral and written communication skills.
Ability to manage multiple projects, keeping oriented to the goals and tracking details, efficiently communicating on issues and outcomes, and executing independently and in collaboration.
Effective organizational, interpersonal, and customer-service skills.
Aptitude for learning quickly and functioning in a dynamic technological environment.

Nice To Haves

Master’s or PhD is preferred.
Experience with parallel computing, GPU operation (CUDA Toolkit), multi-GPU training, and distributed frameworks for machine learning (e.g. PyTorch Distributed).
Prior software engineering experience within AI, scientific computing, or robotics domains in industry, academia, government, or non-profit organizations, which involved architecting and managing codebases for large-scale technical projects
Prior contribution to computationally-focused research publications and/or software, or contribution to creating developer documentation, API guides, and descriptions of software use cases or other write-ups.
Experience with modern collaborative productivity environments (e.g., Slack, Trello, etc.).
Experience utilizing AI-assisted programming environments (e.g., GitHub Copilot, Codex, Cursor, or LLM-driven coding tools) to streamline software development, unit testing, and refactoring.
Course work or experience in one or more areas: AI, data science, machine learning, computer vision, signal processing, knowledge representation (e.g., knowledge graphs), information retrieval, statistics, biophysical modeling, robotics, or neuroengineering.
Experience developing within NVIDIA Omniverse, Isaac Sim, Isaac Gym, or custom CUDA kernels.

Responsibilities

Participate as a member of a highly technical, research driven team, forging close working relationships with HAIC researchers (faculty, post-doctoral, students, and staff) across multidisciplinary working groups.
Provide support to HAIC researchers to use research computing hardware and software for workflows involving AI, data science, machine learning, and physics simulation, especially machine learning frameworks with interfaces for monitoring training, robust checkpointing, and validation suites.
Assist researchers to develop, deploy, scale, and optimize workflows that effectively and efficiently use these high-performance computing resources.
Utilize industry standard research software practices such as establish software version control, pipelines, institute standard testing frameworks, maintain documentation, manage public GitHub repositories to support research dissemination, applying best practices when authoring and assisting in the pursuit of modular and maintainable software codebases that accelerate scientific discovery and societal impact.
Help to identify appropriate hardware, software, and cyberinfrastructure platforms for different stages of research and development, including data collection, storage, and integration, along with computational-focused education and workforce development activities, assist in procurement, set-up, testing, and documentation.
Enable the translation of software and models toward deployment in various settings and platforms.
Serve as a computational liaison to research computing and IT groups at UD and partner institutions.
Educate, train, and advise both novice and experienced researchers regarding software engineering, shared high performance computing environments, associated software and services—adapting to individual needs and experiences—and assist in the development of workshops and other training resources.
Work proactively to broaden engagement between HAIC and the research cyberinfrastructure user community.
Perform other related duties as assigned.
Maintain knowledge of applicable laws and developments in state-of-the-art technology, equipment, and systems to support the institute projects.
Assist in the development and translation of biophysical, kinematic, and musculoskeletal models (e.g., OpenSim, MuJoCo) into the NVIDIA Omniverse environment to support dynamic digital twin simulations.
Implement or accelerate processing and inference algorithms for real-time multimodal data, natural language, signal, and image processing.
Assist in the use and integration of APIs, open-weight models, etc. combined with large graph databases.