LinkedInGitHubLinkMediumEmail

Hi, I'm Claire Schlesinger

I am an experienced undergraduate researcher specializing in machine learning, artificial intelligence, and neural networks. I have contributed to cutting-edge research at Northeastern University and the University of Texas, working on graph neural networks, robotics, and large language models (LLMs) resulting in several papers with over 1000 citations. I am proficient in Python, Java, and Typescript, and have experience with frameworks such as TensorFlow and PyTorch.

Education

Northeastern University, Boston, MA   September 2021 – Present

Khoury College of Computer Sciences                   Expected May 2025

Candidate for a Bachelor of Science in Computer Science       GPA 3.98/4.0

Computer Knowledge

Languages:             Proficient With: Python, Java, C, JavaScript, TypeScript

                                 Familiar With: ARM Assembly, Risc-V Assembly, F#, OCaml

Frameworks:          Torch, Tensorflow, Sklearn, Numpy, Pandas, SciKitLearn, React.js, Node.js

Techniques:           Machine Learning, NLP, LLMs, Data Processing, Data Analysis

IDEs:                        VSCode, Jupyter Notebook, IntelliJ Idea, Eclipse

Software/Tools:      Github, Excel, Word, Vim, Docker

Professional Experience

Data Driven Renewables Research Lab | Boston, Massachusetts      September 2024 – Ongoing

Undergraduate Research Assistant at Northeastern University

  Tested a variety of equivariant graph neural networks to determine their performance on predicting the cleavage energy of slab data.

Participated in a joint research project with the geometric learning lab utilizing generative AI to produce new slab candidates with desired properties.


Autonomous Mobile Robotics Labratory | Austin, Texas    May – August 2024

Undergraduate Research Assistant at University of Texas

      Utilized Code LLMs to generate recovery code for robots when they encountered an error or received new instructions to their task.

      Created a website which allowed a user to generate code and recovery code for a robot.

      Performed case studies to analyze the effectiveness of the recovery system in real world situations.

 

Qimin Yan Group | Boston, Massachusetts      July – December 2023

Undergraduate Research Assistant at Northeastern University

      Developed graph neural networks to predict properties of various crystal structures.

      Utilized spherical harmonic decomposition to improve prediction of elastic tensors through the prediction of spherical harmonic coefficients by steerable equivariant graph neural networks.

      Developed transfer learning techniques to improve prediction of other tensorial properties of crystals.

 

Programming Research Lab | Boston, Massachusetts         January 2023 – Ongoing

Undergraduate Research Assistant at Northeastern University

      Developed a benchmark to test Code LLMs performance on context-based tasks.

      Tested ChatGPT’s ability to write code in a variety of languages to solve HumanEval questions.

      Researched an improved method of fine tuning using a modified form of self-instruct for StarCoder.

      Worked in partnership with UT Austin robotics on a language model for conversion of human commands into robot code.

 

Data Initiative Research Lab | Boston, Massachusetts         September – December 2022

Data Science Consultant at Northeastern University

  Identified features with InnSure Corporation that increase or decrease a company’s susceptibility to be hacked utilizing the Veris Dataset.

  Analyzed data, identified correlations, and presented findings with a team to InnSure management.

 

Cybersecurity and Machine Learning Project | San Antonio, Texas                                  Summer 2019 – 2021

Machine Learning Intern at Denim Group (acquired by Coalfire)

  Constructed and maintained a Python API which handled requests to a ThreadFix instance.

  Created and optimized a classifier that could detect false positives at a 98% F1 Score.

  Built an end-to-end training pipeline for the classifier to work on any ThreadFix instance.

  Developed a website in React.js with an intuitive interface to use the classifier.