Skip to main content

2023 LAS Sponsored NCSU Computer Science Senior Design Projects

Since 2016, researchers at the Laboratory for Analytic Sciences (LAS) have sponsored 18 senior design projects through the NCSU Computer Science Senior Design Center, including five in 2023.

The idea of the Senior Design course is to offer students a hands-on learning experience working on real-life problems sponsored by area businesses and research organizations (the LAS being one of them). Students work in small teams under the guidance/mentorship of representatives from the sponsoring organization throughout the term to develop solutions.

For the LAS, its 18 projects over the years have spanned a variety of technical topics, from cloud computing efficiency upgrades to gamifying the task of machine-learning annotation. All project proposals stem from mission needs in analytic technology and tradecraft from LAS stakeholders, and are developed over the months preceding each semester. After the semester concludes, LAS staff work to transition positive results onto/into mission systems. Significant, tangible successes have resulted. Below we will briefly describe the five projects undertaken this year.

Note: A full accounting of these five projects, or any of the other Senior Design projects that the LAS has sponsored in the past, is beyond the scope of this article. Please see the article we wrote at the LAS 2022 Symposium on the topic, which attempts to provide a more overarching view of the LAS Senior Design engagement (including projects from other departments and colleges), while simultaneously giving brief descriptions of every project throughout the years. For more detail on any specific project, please reach out to the author of this article, or any other POCs noted.

  1. NCSU Computer Science, Spring 2023: “Improving the TLDR Prototype” – Each summer, the LAS hosts a Summer Conference on Applied Data Science (SCADS). SCADS has a 5-10 year Grand Challenge of creating an algorithm that can generate a “Tailored Daily Report”, or “TLDR” for short. A TLDR is meant to be a short report containing a summary of relevant, pertinent information for the user relative to their individual objectives and interests. This Senior Design team took the prototype developed from the previous summer’s SCADS, and added a host of enhancements to it. For instance, new data visualizations, support for multiple concurrent users, and the ability to select amongst available recommendation engines. They also developed detailed User’s, Developer’s, and Installation Guides.
  1. NCSU Computer Science, Spring 2023: “Data Prioritization Manager I” – The team implemented, and successfully tested, a web application designed to enable users to manipulate the handling of data in large systems. In the application, users can create, and edit “rules” and data “buckets”, and associations between them. These then work together to allow the user to effectively manipulate how data from an ingest flow is routed for subsequent processing. It also lets the user control how much data is processed, and with what priority. All this is to address the issue of the ever-increasing amounts of data large organizations need to collect and process.
  1. NCSU Computer Science, Spring 2023: “Using LLMs to Improve STT” – The team created a web application which implements an idea of how a large language model (LLM) might be used to improve a foreign language speech-to-text (STT) algorithm. In some cases, language analysts translate foreign language audio into English text. The concept here is that with this new information in hand (which is unavailable to the STT algorithm), a LLM could be asked to “correct” the original STT output. During testing with a Spanish audio dataset using the team’s newly developed application, this indeed has been the case. The payoff is this provides more “ground truth” data that could be used to improve, or fine-tune, the STT algorithm to generate better results.
  1. NCSU Computer Science, Spring 2023: “Data Prioritization Manager II” – The overarching goal for this team was the same as that described in the “Data Prioritization Manager I” above, as this is a follow-on project with a new team during the subsequent semester. The previous team successfully implemented the basic functionality of the application. This team focused on building out the prototype to include numerous functionality enhancements, as well as a stretch goal of enabling a ML-powered automated prioritization system. The initial prototype requires manual creation of “rules” and “buckets”. This stretch challenge was to start integrating ML algorithms into the application that would automatically generate rules and buckets based off of the ML model’s perception of each individual users’ data needs. This was a very challenging stretch challenge for a single semester, but the team made great progress, along with the many aforementioned functionality enhancements.
  1. NCSU Computer Science, Spring 2023: “Model Deployment Service” – The team implemented, demonstrated, and evaluated the applicability of using the open source orchestration technology Argo Workflows to automate Machine Learning Operations (MLOps) tasks in a Kubernetes cluster. Data scientists are training more operational ML models than ever before, and the challenge of effectively utilizing these powerful models is quickly becoming one of management and orchestration engineering rather than model development. The core intent of this project is to enable automation and scaling in this space, reducing the burden on engineers and data science practitioners. The student team effectively demonstrated that Argo is a powerful tool to help achieve this end goal.