LLM Inference & Integration Engineer Intern

Job ID : 5409

Date Posted : November 4, 2025

The Space Dynamics Laboratory (SDL), a University Affiliated Research Center (UARC), has been developing innovative technologies and solutions for cutting-edge DoD and intelligence programs for over six decades.

SDL’s internship program provides an exciting opportunity for undergraduate and graduate students to get involved with state-of-the art technologies in space-, airborne-, and ground-based systems. With the support of Engineers and mentors, Interns are able to work on professional-level assignments that complement their academic studies. The program also includes training workshops, networking opportunities, and a variety of summer events and activities. Interns will be paid a competitive monthly stipend and will be tasked with varying duties based on current projects, needed support, and development phase.

The Command, Control, Communications, Computers, Intelligence, Surveillance, and Reconnaissance (C4ISR) Systems Division is seeking an exceptional Large Language Model (LLM) Inference & Integration Engineer Intern to assist in building and integrating local LLM inference pipelines.

The C4ISR Systems Division delivers mission-critical solutions, specializing in cyber operations, information architecture, strategic deterrence, and ISR. Our division’s commitment to innovation and security enables us to provide critical solutions across defense, intelligence, and national security. Join our team and contribute to the next generation of defense technologies.

This position has the potential of continued employment or transition to student / full-time employment requiring security clearance eligibility.

This internship is for the summer of 2026.

Primary Responsibilities

Prototypes and productionizes automation workflows that run local, open-weight LLMs using portable inference runtimes and quantized model formats

Integrates LLM chains with internal tools using standard chat / completions-style APIs (prompting, function / tool calling, streaming)

Optimizes inference (quantization choices, context length, batching) for speed / memory on CPU / GPU

Builds small services / CLIs to run batch jobs, evaluate outputs, and log metrics

Writes clear documentation and example notebooks for other Engineers to reuse your pipelines

Contributes tests and lightweight evals to ensure reliability and prevent regressions

Requirements

Experience running local / open-weight models with open-source inference runtimes on CPU and / or GPU

Familiarity with quantized model artifacts / formats and loaders for edge / desktop deployment

Experience calling chat / completions-style REST APIs in code (JSON, streaming, function / tool calling)

Proficiency in at least python, C++, or C#

6+ months of regular development on Linux : Bash scripting, package managers, SSH (key-based auth), Git, file permissions, and basic networking tools; comfortable building from source (CMake / Make) and troubleshooting dependencies

Solid understanding of prompts, context windows, and tokenization basics

Ability to work well independently with minimal supervision

Ability to work well in a team with other students and professionals

Strong initiative and ability to see the job through

Good communication skills, both written and verbal

Must be a US citizen, lawful permanent resident of the US, or other US person

Preferred Skills

Experience with Playwright or similar UI / browser automation

Experience documenting generation pipelines using LLM outputs (Markdown / HTML / PDF, templating)

Basic RAG / evals (embedding stores, quality checks, hallucination guards)

Experience with Docker, CUDA / ROCm, and profiling / benchmarking

Experience with team collaborative tools (Confluence, Jira, GitHub, etc.)

Experience leveraging LLMs to interpret user intent and generate spatial queries over vector and raster data, delivering ranked imagery and regions with documented provenance

Experience applying object detection and segmentation to overhead imagery, including georeferenced tiling, large-scene processing, and packaging inference as scalable batch jobs and services

Education

Must be pursuing a degree in computer science, computer engineering, or electrical engineering

Must be a junior, senior, or graduate student

3.0 GPA minimum is required

If you’re unsure if you are qualified, we still want to hear from you! If you have a passion for problem-solving and learning, you could be a great fit for this internship

SDL delivers advanced multi-domain solutions to protect national security and enable scientific discovery. Our expertise in satellites, sensors and instruments, ground systems and data processing, and autonomous systems plays a critical role in missions supporting NASA and the Department of Defense. Join our team of engineers, scientists, technicians, and business professionals in our seventh decade of delivering mission success.

At SDL, we strive to uphold a culture of respect, collaboration, empowerment, and accountability. We listen with open minds, seek to understand diverse perspectives, and engage in thoughtful dialogue. We work together by sharing knowledge, involving others, and offering support. We trust and empower our team members to take ownership, act with integrity, and be accountable. Above all, we deliver on our commitments to each other and to our mission partners.

The application window for this position is expected to remain open for approximately 7 days. However, it may be shortened or extended depending on business needs and the availability of qualified candidates. We encourage interested candidates to submit their applications promptly.

For questions, assistance, or accommodation with the application process or the DoD SkillBridge program, please contact employment@sdl.usu.edu.

Powered by JazzHR