Summer Internship
Summer Internship
2026 WINLAB Summer Internship
The WINLAB Summer Internship Program offers full and part-time summer internships in a university research setting to highly talented undergraduate and graduate students. The main goal of the program is to provide students with a real-world, team-based research experience in various topics related to wireless technologies. Each intern joins an active research group consisting of a mix of graduate and undergraduate students with at least one mentor who is a faculty member. All projects are designed to be completed within the duration of the program, but can also be extended for eligible students to the following academic year. Each week students are expected to report on the progress of their work in a summer research group meeting. At the conclusion of the program, interns submit a report and are required to give a final presentation on the research results. A limited number of full-time internship students receive a monthly stipend plus an on-campus room in designated Rutgers dormitories (available for non-Rutgers, full-time interns ONLY!). Additional students will be offered part-time (hourly) summer employment. The opportunity for non-paid participation may also exist once the paid positions are distributed. Should you be interested in one of those positions if not chosen for a paid position, please make sure to indicate that on your application. The program for 2026 will begin on Tuesday, May 26th and end on Thursday, August 7th. Please note that the WINLAB facility is NOT on any of the Rutgers bus routes and that getting to/from the lab requires personal transportation.
To apply for the 2026 WINLAB Summer Internship Program, students must be currently enrolled full time in a college or university, be eligible to work in the US, have an anticipated graduation date of 2027 or later, and complete the following five steps:
- Obtain a copy of your transcript. If you are a Rutgers student, an unofficial copy is sufficient.
- Please obtain permission from two recommenders (typically faculty at your home institution or past job supervisors) who can assess the quality of your academic performance and research potential. You will list their names and email addresses in the reference section of the application form, and the system will automatically email them instructions for submitting their letters. If you are a Rutgers student and/or you plan to use a WINLAB professor as a reference, you do not need to request a separate letter from them. Simply list the professor’s name and email in the reference section of the application; WINLAB faculty will be asked for input directly for any student who lists them as a reference.
- Write a brief essay (no more than one page) on why you would like to join the program, what strengths you will bring to the program and what you hope to achieve by being included in the program. Please see the list of research topics at the bottom of this page and advise which projects peak your interest. Please understand that we cannot promise that you will be assigned to the project you ask for, however, we will make an effort to put accepted students in their areas of interest.
- Prepare a CV/resume.
- Complete the application form (requires Google/ScarletMail account) with the above transcript, essay and CV uploaded no later than April 12th. Due to the number of applications that we anticipate receiving this year, incomplete applications or those that are received after the deadline may not be considered. The selection of interns will be determined by the WINLAB faculty members. All accepted undergraduate and graduate students will be notified by email of their acceptance into the program by April 27th
2026 Research Projects
| Project | Pre-requisites |
|---|---|
| mmWave-based passive human activity recognition for smart homes: This project explores using mmWave to perform human activity recognition for smart home applications | Some background in wireless communications and basic programming skills |
| Evo-Driven Protein Variant Generation and Structure Validation: Explore an information-processing problem centered on the protein databank and connect it to Arc Evo 2. The project will select a focused protein group (e.g., enzymes used in biomanufacturing), use Evo to generate sequence variants, and then apply folding algorithms (e.g., AlphaFold3 or ESMFold) to predict 3D structures. Predicted structures will be compared against “gold standard” structures from the Rutgers PDB, leveraging relevant work from the earlier PDB-focused summer effort. | Linux/CLI + Git (working on servers, managing runs, version control). HoloLens / Mixed Reality development basics; Unity + C# (strongly preferred). Python proficiency (data wrangling, scripting, running pipelines) |
| Evaluation and Mitigation of Machine-Induced Noise on SDR-Based OFDM Signal Reception in FR2/FR3 Industrial Spectrum: Investigate how electromagnetic noise generated by industrial machines affects OFDM signal reception in FR2/FR3 frequency bands and evaluate signal processing techniques to mitigate its impact using software-defined radios. | Basic knowledge of wireless communications, MATLAB/Python programming; familiarity with software-defined radio platforms (e.g., GNU Radio or USRP) is helpful. |
| Multimodal AutoLabeling at Scale: Build a scalable multimodal data labeling pipeline for a representative dataset. The project includes engineering loaders for video, sensor streams, and metadata; aligning data across time and modalities; and implementing batch labeling workflows with caching, retries, uncertainty scoring, and failure tagging. It also covers defining an initial label ontology and task schema, creating a human-validated “gold” subset, and adding basic quality-control checks and uncertainty analysis. The summer deliverable is a working end-to-end pipeline with initial QC so the broader program can operate on reliable labeled data. | Strong Python programming and familiarity with ML/data pipelines. Ability to work with video or multimodal data. Willingness to integrate labeling models/APIs and to do ontology design and basic QC |
| Magic Room: Device-free Sensing of People using batteryless tags and machine learning: Variations in multipath signal propagation within a room provide useful information for estimating the number of people present and tracking their movement without requiring them to carry any devices. Batteryless, ultra-low-cost RFID tags can act as passive antennas, significantly increasing the number of measurements in a cost-effective way. Machine learning methods will be used to analyze these signals, with transfer learning techniques explored to improve performance across different rooms. | Software: Python, Pytorch, Matlab |
| Channel Measurement Campaign for Data Analytics in Wireless Networks: This project will develop an SDR based platform for channel characterization and coverage measurements on COSMOS/ORBIT testbeds and perform a series of measurement campaigns. Once the platform is developed, a series of measurement campaigns will need to be performed at various locations to validate the platform’s performance and accuracy. These campaigns will involve deploying the platform at different locations on the COSMOS/ORBIT testbeds, capturing wireless signals from different sources, and analyzing the data to derive insights into the wireless channel characteristics and coverage. | OS: Linux Software: C/C++, Python, GnuRadio, PyTorch, TensorFlow |
| Self-Driving Vehicular Project: This project will assemble and train miniature autonomous vehicles to run in the miniature smart city environment, by using low latency networks for vehicular control. This project will use specialized low latency cameras and radios to operate remote model cars and design and implement self-driving algorithms using machine learning libraries in python. Students will design behavior that will allow the vehicles to react realistically to other cars and props in the smart city environment, and work with the testbed infrastructure to use external data from the intersection to improve performance. Another goal is to offload localization, mapping and navigation to MEC using 5G and get back commands to control the wehicle/robot. | OS: Linux, RoS Software: C/C++, Pytorch, Tensorflow |
| Digital Twin-Guided OTFS vs. OFDM Waveform Selection for Outdoor NextG Communications: This project builds a site-specific digital twin of the WINLAB outdoor scene using OpenStreetMap and Sionna ray tracing to generate delay-Doppler channels under varying mobility (pedestrian, cyclist, vehicle). OFDM and OTFS performance are compared across positions and speeds to produce a waveform selection map, validated against outdoor SDR measurements at sub-6 GHz/mmWave. We can see in which velocity OTFS does better and if we could make some reliable predictions. | Requirements: Linux, Python, Sionna, 3D Modeling, GNU Radio, USRP B205mini. Background in communication systems, signal processing, and Python programming. Familiarity with OFDM preferred. |
| O-RAN network energy management: In this project, we extend the management capabilities of the WINLAB open-source O-RAN network testbed. The project involves implementing/improving the M-plane functions to control simulated and real O-RAN Radio Units to research and demonstrate energy management | Graduate student or advanced undergraduate with relevant experience. Knowledge of Linux/Python required. Knowledge of netconf-based management, OAI/srsRAN open-source would be an advantage |
| Smart Room as an Indoor CityOS Node: This project builds an indoor CityOS node with two students: one focuses on runtime and integration; the other on data collection and evaluation. The smart room is an indoor CityOS deployment (API 1 style ephemeral sensing, local aggregation, privacy-aware release). Together they deliver a working sensing node, a curated dataset slice, evaluation, and a live demo for TeLLMe and the labeling pipeline. | Solid Python skills, sensors/video/event pipelines, Linux |
| O-RAN Curriculum Development: Create comprehensive educational resources, including modules, test scenarios, and hands-on lab exercises, focused on equipping aspiring engineers with the skills needed for O-RAN testing and validation. Collaborate experienced researchers, delve into O-RAN standards, design practical testing scenarios, and build a resource library, ultimately contributing to a framework for evaluating O-RAN testing expertise and exploring automated testing methodologies. | OS: Windows, Linux Basic RF Measurements |
| Efficient AI and Backscatter Radio Tags for Interference Mitigation: This project focuses on the design and construction of backscatter radio tags that function as a reconfigurable intelligent surface (RIS) to mitigate wireless interference. By controlling how the tags reflect radio signals, the system aims to protect a receiver from interfering signals and improve communication reliability. AI techniques will be explored to optimize the configuration and operation of the tags for effective interference suppression. | Embedded programming (in any language or platform), Python, Pytorch |
| AR Mural: This project involves developing an augmented reality (AR) based art platform that allows users to contribute their artwork, including paintings, photos, videos, and sculptures, to a set of locations. The 3D sculpture building feature will enable users to create and contribute to 3D sculptures that will be shown simultaneously across multiple locations. The team will need to develop the necessary tools and algorithms to enable collaborative sculpting, as well as implement the required backend infrastructure and communication protocols to ensure that the art can be displayed in real-time across multiple locations. | OS: Windows, Linux Software: Unity, C#, C/C++ |
| CARLA Simulated User Interaction for CityOS: Build a simulated user agent that interacts with a CityOS-like workflow inside CARLA, run experiments, and document a transition path to the lab. The project uses an LLM-based agent to generate queries and commands in a virtual urban environment, with coordination and limited diagnostic integration (e.g., ConcordFS, TraceFix) included as feasible. | Experience with simulation frameworks (e.g., CARLA) or strong willingness to learn quickly. Strong Python skils and familiarity with AI/ML and LLM integration. Ability to design and implement an agent and connect it to a workflow. Ability to run experiments and document a transition plan |
| Core Infrastructure for CityOS: ConcordFS, TraceFix, and Verified: This project combines the coordination substrate (ConcordFS), trace analysis and diagnosis (TraceFix), and verified coordination with runtime enforcement. The three students work as a mini-team: ConcordFS provides durable coordination and logs; TraceFix consumes those logs for debugging and recovery; Verified Coordination model-checks selected workflows and adds runtime safeguards. Together they deliver the core infrastructure that the smart room, labeling, TeLLMe, and CARLA projects depend on. | Strong programming skills, Python, systems thinking |
| Coordination: The size and timing of packets sent by devices can leak information: how can we “shape” the traffic to enhance privacy? | Basic probability, Python and PyTorch |
| Virtual reality enabled activity recognition for immersive computing: This project explores sensing in virtual reality headsets to perform activity recognition for immersive computing. | Some background in virtual reality and basic programming skills |
| Green Communications: Today’s cell towers are large, complex and expensive, but recent advances in thin-film solar and lithium iron phosphate (LiFePo) energy storage will allow for a new class of small-scale communication infrastructure. In this project, you will use a combination of thin film solar panels, batteries and Long Range (LoRA) wireless radios to build a small scale demonstration sensing and communications platform with a tree-like structure. The platform will use a combination of solar power, energy storage, and a small neural network to provide image recognition, wireless mesh networking, and small device charging services. | OS: Linux Software: Python, C#, C/C++ |
| Visual Perception for AR Glasses: Augmented Reality (AR) requires continuous processing of visual data to enable low-latency interaction with the environment. This project focuses on enhancing the visual perception performance of AR glasses. The main goal is to develop an efficient visual perception system that leverages computation offloading to improve processing speed and reduce power consumption. Students will explore and implement advanced computer vision techniques while optimizing the system for low-latency execution, ensuring smooth AR experiences. | OS: Linux Software: Python, Go, C/C++ |
| Quantum- and Physics-Inspired Computing for Large-Scale MIMO Systems in NextG Wireless: The project will explore various non-traditional Post-Moore computing methods to accelerate computationally intensive processing in MIMO systems, including, but not limited to, uplink MIMO/NOMA detection, downlink MIMO precoding, and LDPC decoding | OS: Linux, Windows Software: Python, C/C++ |
| Physics-Inspired Restricted Boltzmann Machine Processing for Real-Time Wi-Fi Human Activity Recognition: The project will couple physics-inspired computing with machine learning training for real-time wireless sensing applications. | OS: Linux, Windows Software: Python, C/C++ |
| Minimal Neural Network Structures for Spatial Problem Solving: A jumping spider, possessed with a paltry number of neurons, can observe a maze from above and then show improved navigation when placed inside. This kind of spatial problem solving, where learning occurs from a different perspective than actuation, must therefore be "simple," yet it is something that we struggle to replicate in our autonomous systems. This project asks, what is the minimal artificial neural network structure required to replicate the jumping spider’s ability? | OS: Linux Software: Python, Pytorch Preferred: proficient with Python, some experience with machine learning |
| TeLLMe: Grounded Query Interface: Develop a user-facing query interface on top of the CityOS platform. The interface (TeLLMe) answers grounded questions about smart-room events, CityOS-style aggregates, labeling outputs, and—where useful—ConcordFS traces. Responses are evidence-linked, pointing to specific events or traces rather than only freeform text. The summer deliverable is a polished demo in which a user asks questions about the environment or data and receives grounded, evidence-backed results. | Solid Python (or equivalent) and ability to build a simple query interface. Interest in LLMs, retrieval, or question-answering systems. Ability to connect to APIs and structured data sources. Willingness to iterate on demos and user experience |
| AI for Wireless: Deep Learning-Based CSI Feedback and Precoder Design for 5G MU-MIMO: This project investigates an end-to-end AI-enabled framework for 5G MU-MIMO communication in which deep learning is used for channel state information (CSI) compression at the UE side and CSI reconstruction at the gNB, followed by AI-assisted precoder design at the gNB for energy-efficient downlink transmission while satisfying user quality-of-service (QoS) and rate constraints. | OS: Linux or Windows Software: MATLAB Background: Basic communication systems, signal processing, and programming Preferred: Familiarity with matrix computations, neural networks, and machine learning or deep learning concepts |
| Benchmarking Information Processing Methods for Edge-Local Data Stations: This project develops a reproducible benchmark framework to evaluate how different information processing methods perform when deployed on resource-constrained edge devices. The intern will build a data ingestion pipeline that collects information from multiple public sources (e.g., news feeds, academic APIs, sensor streams), then implement and compare 3–4 processing approaches—rule-based filtering, ML-assisted extraction, LLM-based summarization, and agentic retrieval—measuring latency, accuracy, and computational cost on standardized datasets with curated ground truth. The framework will be containerized for portability and designed so that each experiment is fully reproducible with a single command. The deliverable is an open-source benchmark suite with a detailed comparison report, providing empirical guidance on method selection for edge information processing scenarios. | OS: Linux Software: Python, C#, C/C++ |
| Summer Internship Helper (graduate students only): Assist in the day-to-day coordination and mentoring of summer interns participating in the WINLAB research internship program. Responsibilities include helping interns set up development environments, troubleshooting technical issues, guiding them through research methodologies, and supporting their final project presentations. The helper will serve as the primary point of contact for interns, facilitating communication between interns and their faculty mentors. | Prior TA experience. Familiarity with Python, MATLAB, PyTorch, and TensorFlow. |
| Embedded machine learning models: Recent years have shown great strides in machine learning models. In this project you will build a deep neural network model that runs inside networking cards to classify network traffic. The model is pretrained using existing software. The goal of the project is to write the pretrained model in a hardware description language such that it can run on specialized networking infrastructure at networking speeds. The first third of the project will consist of learning a hardware description given knowledge of a general purpose programming language. The remainder of the intership will develop the model to run on line-rate networking cards. At the end of the project the neural network should be able to classify traffic at multi-gigabit speeds. | Programming and software development experience in Python, Java or C\ |
| World Cup scaling for 100 million viewers with 5G MBS: How do we prevent 5G networks from collapsing when 100 million fans stream a Cricket or Soccer World Cup Final simultaneously? This project tackles the "Concurrency Crisis" using 3GPP Release 17 Multicast-Broadcast Services (MBS). The intern will implement a "Transparent Mode" testbed using 5G-MAG and srsRAN to demonstrate high bandwidth, energy, and spectral efficacy compared to traditional Unicast delivery. | High-level Linux systems proficiency (Networking, UDP/Multicast, IP tables). Ability to compile and modify complex open-source C/C++ projects (srsRAN/Open5GS). Expert-level packet analysis (Wireshark/tcpdump). Preferred: Previous experience with Ettus/USRP SDRs. A Linux "hacker" mindset. |
| Resilient text messaging networks: Recent world events have shown the need for resilient networks that can operate when existing networks are compromised, shut down, or generally unavailable. In this project you will develop a small solar powered networking relay using Long Range (LoRA) radios. The relays will be powered by an unusual design of pole-mounted, small vertically oriented thin-film solar panels. The array of small panels can be robust to damage, tolerate shading, and power the relay through ice and snow. The relays will run Meshtastic, which is an open-source, decentralized, and encrypted peer-to-peer (P2P) mesh networking protocol designed for off-grid communication. The goal of the project is to show a text-messaging mesh with a several kilometer range. | Programming and software development experience in Python, Java or C. Basic Circuit Design |