CTL 2026 SURF Program

Internships with the Communications Technology Laboratory

The 2026 SURF program is designed to provide a paid 11-week hands-on research experience in communications technology. CTL's SURF projects offer virtual and in-person opportunities in Boulder, Colorado, and Gaithersburg, Maryland.  Interns are awarded up to a $7,810 stipend, with travel and housing allowances available for those who live outside of the NIST immediate area. Applications are due January 26, 2026. 

With expertise honed over decades of research in antennas and wireless propagation, materials science and electronics testing, as well as communications network protocols and standards, CTL serves as an independent, unbiased arbiter of trusted measurements and standards to government and industry. We focus on developing precision instrumentation and creating test protocols, models, and simulation tools to enable a range of emerging wireless technologies. Click here for additional information on CTL.  

Opportunities located in Boulder:  CLICK HERE TO APPLY 

RF Technology Division (Read More)

Opportunities in Gaithersburg  CLICK HERE TO APPLY 

Wireless Networks Division (Read More)

673-1 Simulation and Modeling of Resilient Mobile Networks

Wesley Garey, 301-975-5190, wesley.garey [at] nist.gov (wesley[dot]garey[at]nist[dot]gov)

Wireless mobile networks have become an integral part of everyday life as cellphones have become ubiquitous in today’s world. These gains have taken place over several decades with advent of 3G which enabled data communication, 4G that enabled broadband connectivity, and 5G that is now capable of providing ultra-fast speeds and ultra-low latency. The latest enhancements to 5G and moving forward with 6G will provide several features with the intended purpose of extending network coverage and maintaining user connectivity. This includes the use of Non-Terrestrial Networks (NTN), i.e., satellites, and direct device-to-device (D2D) communication. The NIST Wireless Networks Division (WND) is actively developing next-generation cellular communications standards and performing research studies around network resiliency and ubiquitous connectivity via the use of NTN and D2D communication.

The student selected for this project will have the opportunity to work alongside members of WND to create and enhance software tools used to evaluate wireless communication technologies. This includes developing visualization tools to demonstrate simulation events, utilities to facilitate Continuous Integration and Deployment (CI/CD) of the developed tools, and simulation models in ns-3 and other platforms to simulate emerging wireless technologies. The student selected for this project will code using a variety of programming languages, regularly use the Linux command line, and manage code using a version control system such as Git.

Desired Skills/Experience

Major in Electrical Engineering, Computer Science, or a related field with some programming experience. Knowledge of C++, Python, and/or Bash is strongly recommended.  Background in wireless networking (principles of the Internet and cellular communications).  Experience using Git.  Interests in computer models and simulations of future networks.  Ability to collaborate with a team and good communication skills.

673-2 ML Sensing Abstraction for  ISAC System Level Simulation

Anirudha Sahoo and Steve Blandino 301-975-4439, anirudha.sahoo [at] nist.gov (anirudha[dot]sahoo[at]nist[dot]gov)

Future wireless networks won't just transmit data; they will use radio waves to sense the physical world, acting like a radar to detect objects and their locations. This is known as Integrated Sensing and Communications (ISAC).

To design these systems, researchers run complex simulations that model exactly how radio waves bounce off objects. These simulations are extremely accurate, but they are also incredibly slow and computationally heavy. It is currently impractical to run these simulations at a large scale to test entire networks.

Instead of calculating the physics of every radio wave, we want to build a Machine Learning model that can "predict" the outcome instantly. In this project, the student will build a fast "surrogate model" (an AI substitute). The student will train this model to look at the geometry of a scene, e.g., where the base station is, where the target is, and how many antennas are used, and accurately predict how well the sensing will work. The objective is to teach an AI to replicate the results of a complex physics engine, but in a fraction of the time.

Desired skills/experience

Essential skills include a solid understanding of Machine Learning (ML) concepts and model training, proficiency in Python or MATLAB, and a good understanding of Linear Algebra. Preferred skills include understanding the basics of signal processing concepts and familiarity with wireless / Radio Frequency (RF) concepts such as, Signal-to-Noise Ratio (SNR), Orthogonal Frequency-Division Multiplexing (OFDM), and antenna arrays.

673-3 Enabling 6G: Performance Benchmarking of ISAC via 3GPP Standardized Channels

Jian Wang and Neeraj Varshney, 301-975-8012, jian.wang [at] nist.gov (jian[dot]wang[at]nist[dot]gov)

Integrated sensing and communication (ISAC) is set to become a key advancement in next-generation wireless systems, transforming the widely deployed communication network into a sensing-enabled infrastructure. Future wireless systems are expected to support sensing capabilities alongside communication, enabling devices to detect and track objects such as UAVs, humans, and industrial vehicles (AGVs) using radio signals. Simulation is a powerful tool for researching and evaluating ISAC systems, which consist of several components that simulate a wireless signal transmitted through the wireless channel and received at the receiver, then further processed to detect targets. Among these components, the channel model enables a realistic simulation of the wireless propagation environment and models realistic channel effects.

To evaluate the performance of ISAC, 3GPP, a leading global organization in cellular standardization, has recently developed an ISAC channel model that includes target reflections, clutter, and geometry-dependent propagation effects. Although the implementation for UAV sensing has already been completed, many important use cases, such as human sensing and AGVs operating inside factories, still need to be developed and tested. 

The student will collaborate with a group of researchers to extend the current ISAC channel model implementation and introduce the new sensing targets, considering the 3GPP deployment recommendations. The student may later use the 3GPP-based channel model to evaluate sensing performance. The project bridges channel modeling, sensing signal processing, and system-level performance analysis, providing hands-on experience with standards-driven ISAC research.

Desired Skills/Experience

Major in Electrical Engineering, Computer Science, or a related field; proficiency with MATLAB; experience in digital signal processing; familiarity with Git version control; solid mathematical foundations in probability and matrix operations; and strong communication skills.

673-4 Prototype and Evaluation of Security Protocols for 5G/6G Mobile Networks

Scott Rose and Oliver Borchert, 301-975-8439, scott.rose [at] nist.gov (scott[dot]rose[at]nist[dot]gov)

5G/6G Open-Radio Access Networks (O-RAN) technologies aim to transform radio access networks from single-vendor solutions based on proprietary appliances to a disaggregated network architecture of components and functions, featuring standardized open interfaces designed for deployment in virtualized and cloud-native environments. NIST is actively engaged in O-RAN Alliance standards development, focusing on enhancing the security of virtualized, cloud-native O-RAN functions. We see this area as having both the most significant potential to increase overall network security and the most significant potential risk to the eventual commercial viability of O-RAN technologies. Some of these new technologies have not been designed with mobile networks in mind, but may help secure next-generation telco networks. This project will involve prototyping new protocols and methods of providing secure communication between virtualized workloads. This involves working on the overall design, development, deployment, and testing in a cloud-native environment. Knowledge of virtualization platforms, such as Kubernetes, is helpful but not necessary.

Desired skills/experience

Linux, Kubernetes / Docker, service-based architectures, some programming (Python, Golang, NodeJS, or similar), dev-ops / network programming, network protocols/tools/technologies (HTTP, TLS, PKI, OAUTH, Wireshark), security scanning tools.

673-5 Building Edge-AI Smart Sensor Testbed for ISAC

Jian Wang, 301-975-8012, jian.wang [at] nist.gov (jian[dot]wang[at]nist[dot]gov)

Integrated sensing and communication (ISAC) is poised to become an intrinsic feature of the next-generation wireless systems. In addition to leveraging the same waveform for simultaneous sensing and communication to achieve good spectrum/hardware efficiency, the environmental information collected through sensors can also enhance communication and improve its performance, thereby enabling sensing-assisted communication. For example, continuously updated environmental information can be used to predict link blockage and mobility, enabling the network to utilize this information for beam selection, predictive handover, and interference management to maintain reliable connections.

Sensing data can be collected through various types of sensors, including radar, Camera, and LiDAR. Further processing is often required to understand the radio environment, including the mobility of objects within it, their velocities, motion patterns, and other relevant factors. To reduce sensing latency, conserve bandwidth, and enhance privacy, Edge computing is a vital solution by processing data close to where it is generated, often on local sensors or gateways, rather than relying on the cloud. This project will focus on developing AI-enabled smart sensors, building a testbed with Raspberry Pi and Camera/Radar sensors, creating software tools for object detection using AI/ML algorithms, and writing a technical report to document the work. This project will be conducted in collaboration with a team of NIST researchers.

Desired Skills/Experience

Major in Electrical Engineering, Computer Science, or a related field; proficiency with Python; experience in AI/ML and Edge AI; familiarity with Git version control; and strong communication skills.

673-6 Automated Compliance Assessment of Scientific Manuscripts Using LLMs

Richard Rouil, 301-975-3387, richard.rouil [at] nist.gov (richard[dot]rouil[at]nist[dot]gov)

This project aims to accelerate the manuscript review process by developing an LLM-based framework to assess draft manuscripts for compliance with publication guidelines. The student will engineer a system capable of identifying specific technical requirements, such as the proper application of SI unit standards, mandated uncertainty statements, and the appropriate use of commercial disclaimers. The scope involves implementing a prompt-based or Retrieval-Augmented Generation (RAG) pipeline to ground the model in current NIST policies and conducting comparative evaluations of various solutions. By benchmarking AI-generated feedback against historical reviews, the student will quantify the system's accuracy and reliability.

Desired Skills/Experience

Experience AI/LLM. Computer/Data science major is preferred.

Smart Connected Systems Division (Div. 674) (Read More)

674-1 Testable Cyberphysical System Specifications

Charles Manion, 301-975-4251, charles.manion [at] nist.gov (charles[dot]manion[at]nist[dot]gov)

This project provides an opportunity for a student to learn systems modelling in the new Systems Modeling Language v2 (SysML2) standard by developing modelling methods to describe and test spatiotemporal behavior. SysML is a widely-used standard for describing complex systems, such as telecommunication networks, spacecraft, naval vessels, and manufacturing systems, enabling large engineering teams and AI to collaborate in designing and analyzing them. SysML v2 adds new spatiotemporal modeling capabilities that enables intended behavior to be described in a much more detailed and testable manner.

SysPhS adds 1D modeling to SysML and defines translation to 1D simulation tools, such as OpenModelica and Mathworks Simulink/Simscape.  This kind of modeling assembles physical and signal components that include ordinary differential and differential algebraic equations forming a system of equations solved by time-stepped simulators.  It is applicable to a wide variety of cyberphysical systems, including signal processors, electrical, mechanical, hydraulic, and thermal. This project will focus on investigating how to specify intended behavior in SysML2 and test that systems simulations have said intended behavior.

The student will (A) Learn how to model in SysML 2. (B) Model a system in SysML2 and investigate techniques to test its behavior (C) Develop new physical/signal component libraries for SysPhS in SysML 2.

Desired skills/experience

Preferred major: Mechanical, Aerospace, or Electrical Engineering. Required: Calculus based physics or dynamics or circuit theory. Basic programming experience such as with Python or MATLAB. Recommended: Experience with 1D modeling and simulation such as Modelica or Simulink. Has taken Numerical methods, linear systems and signals, or controls. Some familiarity with object-oriented programming.

674-2 V2X Network Performance Assessment Tool

Eugene Song, 301-975-6542, eugene.song [at] nist.gov (eugene[dot]song[at]nist[dot]gov)

Vehicle-to-everything (V2X) communication, a critical component of Intelligent Transportation Systems (ITS), is a wireless technology that enables vehicles to communicate with their surroundings (e.g. other vehicles, roadway infrastructure like traffic lights, and pedestrians) to improve road safety and traffic efficiency of safety-critical applications (e.g., emergency brake warning, intersection collision warning). However, one of main challenges of V2X communication for these safety-critical applications is network performances (e.g., latency, reliability, and bandwidth utilization) of V2X devices (e.g., on-board units (OBUs) inside vehicles and road-side units (RSUs) installed along the road and at the intersection). Therefore, V2X network performance testing, measurement, and assessment methodologies are keys to overcome this challenge.

This project will focus on developing performance analysis methodology and software tool of V2X network communication, including design and develop an open-source software tool to analyze network performance of V2X devices. The student will study performance testing methods, measurement and assessment metrics/methods, design and develop an open-source software tool to automatically analyze network performance of V2X devices, conduct performance analysis using the software tool developed and network packet datasets collected in performance testing experiment, write a technical report including software tool design and performance analysis results, prepare and provide a SURF presentation. This work will be performed collaboratively with a team of NIST researchers.

Desired Skills/Experience

Java or Python programming experience. Computer Engineer or Computer Science major preferred. Experience with Extensible Markup Language (XML) and GitHub.

674-3 Understanding Communication Performance of V2X (Vehicle to Everything) Messages Over Satellite Channels

Wendy Guo, 301-975-5855, wenqi.guo [at] nist.gov (wenqi[dot]guo[at]nist[dot]gov)

This project defines and executes a simulation-based study to evaluate the performance of Vehicle-to-Everything (V2X) messages when transmitted through Non-Terrestrial Network (NTN) channels, specifically utilizing Low-Earth Orbit (LEO) satellite links. Employing established tools like MATLAB or ns-3, the methodology involves generating representative V2X traffic (e.g., Basic Safety Messages and MAP/SPaT) and modeling key NTN impairments, such as significant propagation delay and packet loss. The performance is systematically compared against an established terrestrial communication baseline, quantifying the trade-offs and viability of using LEO satellite links for vehicular communication.

The evaluation goes beyond conventional measures of communication reliability (packet delivery rate, latency) to include the critical application-layer metric of semantic usefulness. This analysis determines the operational relevance of delayed or corrupted messages by assessing their information freshness and suitability for real-time vehicular decision-making. The student will analyze how NTN-induced impairments affect different layers of V2X interoperability, with the final deliverable being a comprehensive technical report and presentation detailing the quantitative results and findings.

Desired Skills/Experience

Computer Science or Network Engineering major preferred. Proficiency in MATLAB, Python for data analysis and simulation. Strong analytical skills and an interest in communication systems and simulation-based research.

674-4 Analysis of Prompt Variation on LLM-generated Data Quality

Michael Dawson, 301-975-0671, michael.dawson [at] nist.gov (michael[dot]dawson[at]nist[dot]gov)

Conversational interaction with generative artificial intelligence models has become a popular method for interacting with collections of information; a user sends a prompt to the model, and the model provides a response. These models can be prompted to generate synthetic data, which is used to supplement real-world datasets containing few samples. The most important requirement for synthetic data is that it accurately emulates the data it is meant to supplement. If it doesn't, any tool or analysis based on that data is invalidated. In this project, the SURF student will evaluate the relationship between generative AI model prompts and the quality of the corresponding output synthetic data. Statistical tests will be used to compare the generated datasets to a reference, non-synthetic, tabular dataset. Prompt parameterization will be used to create inputs to the model programmatically, in order to study the effect of small variations in prompts. Three types of prompts will be evaluated: (i) textual description of the reference dataset, (ii) textual description of the reference dataset and details on each of its columns, and (iii) a copy of the reference dataset. The student will learn about classical natural language processing, generative artificial intelligence, and statistical methods.

Desired Skills/Experience

Interested students must have experience using Pyhton and Git. Knowledge of introductory statistical concepts is preferred.