Samar
Abu Hegly

I research and build at the intersection of people and technology, starting from the gap between what a system was built for and how it serves the people using it, with a focus on making AI technologies more human, ethical, and accessible.

Get in touch

01 โ€” ABOUT

Marhaba ๐Ÿ‘‹ู…ุฑุญุจุง ๐Ÿ‘‹

I'm Samar (pronounced similar to summer). I studied Computation and Cognition at MIT, where I completed both my bachelor's and master's degrees. Currently, I am working across several research projects, teaching, and volunteering at my local community center, where I help bring technology literacy and skills to the community.

I've always believed technology can change the world for the better, but the harder question, the one I've spent years trying to answer, is how I can be part of making that change. Exploring across fields, I kept arriving at the same place: the most positively impactful technologies are the ones designed as tools in service of people, not as ends in themselves. That's what drew me to human-centered technologies.

I'm a researcher at heart, curious and willing to follow a problem wherever it leads. I'm also a builder. I think in systems and prototypes and want to see ideas become real products. And I'm an educator, because teaching has been one of the most grounding parts of my journey; building with and for people means meeting them where they are. I work best across disciplines, and I believe the people a system is meant to serve should have a voice in shaping it, whenever that's possible.

When I'm not at my desk, you'll find me on long walks by the water, spoiling my dog, or deep in a book (I'm always open to new recommendations!).

Samar Abu Hegly smiling outdoors, holding her dog Mario

02 โ€” RESEARCH

Research

HCIAI / MLEducationAccessibility

Technology rarely works the same way for everyone, and understanding why and what to do about it is the underlying thread that connects my work. I approach problems from multiple angles, combining design, evaluation, and implementation, and drawing on whatever disciplines the question demands. That's taken me across AI, education, and accessibility: studying how tools perform in real classrooms, building systems that work for students they were never designed for, and asking whether AI-powered technologies actually deliver on what they promise. I'm interested in the full arc from research question to working system, and in the spaces where those two things don't yet meet.

M.ENG THESIS

AccessibilityAI Education

Blocks Without Barriers: Making Blockly Accessible for Blind and Visually Impaired Students

2024 โ€“ 2025ยทMIT RAICA Lab

Block-based programming environments are commonly used to introduce programming and AI concepts, but are often inaccessible to blind and visually impaired learners. This thesis explores how a Blockly-based environment can be adapted to help these students program independently. Two portable, co-designed plugins were developed: a custom screen reader providing real-time audio feedback using the Web Speech API, and an enhanced keyboard navigation plugin extended with student-tested features. Development followed an iterative, user-centered co-design process, with feedback from blind and visually impaired students directly shaping implementation priorities. By the final session, all participants were able to complete essential tasks to build functional programs independently.

PUBLICATIONS & ACTIVITIES

AI EducationAAAI 2025

Advancing Research on Equitable AI Education Through a Focus on Implementation

Bosch, C. A., Gustafson-Quiett, M. C., Abu Hegly, S., et al.

Insights from a middle school computer vision module beta-test, examining how equitable AI education translates from research to real classroom implementation.

Read paper โ†’
AI LiteracyAAAI 2025

Supporting AI Literacy Teaching Through the Development of Assessments for Classroom Use

Masla, J., Bosch, C., Ravi, P., Abu Hegly, S., et al.

Developing and validating classroom assessments that help teachers gauge AI literacy in their students, bridging research and practice.

Read paper โ†’
AI EducationAAAI 2025

Act Out An LLM

Wharton, S., Masla, J., Guterman, L., Gustafson-Quiett, M. C., Bosch, C., Abu Hegly, S., et al.

An activity in which students act as a large language model to generate a sentence based on a prompt. Students are introduced to the concept of a token in text generation through a next-token-voting activity repeated until a full sentence is generated.

View activity โ†’

TALKS & PRESENTATIONS

AccessibilityCATE 2025

Lessons Learned in The Path to Accessibility: Enhancing the RAISE Playground for All

Presented findings and lessons learned from the process of making the RAISE Playground accessible to blind and low-vision students.

View slides โ†’

RESEARCH POSITIONS

AIRE Lab

Research Assistant

Apr 2025 โ€“ Present
  • Currently leading the design of an AIRE Lab pilot study examining whether an LLM-integrated assignment methodology strengthens critical thinking, verification behavior, and AI literacy among Arabic-speaking middle school students.
  • Contributed to the evaluation of a patented LLM-integrated assignment methodology designed to preserve students' higher-order and critical thinking skills; analysis-informed assessment design and implementation insights.
  • Developed a qualitative codebook and conducted thematic coding of 100+ student artifacts; supported inter-rater reliability checks in collaboration with the research team.
MIT RAICA Lab

Graduate Research Assistant

Jan 2024 โ€“ Aug 2025
  • Designed and shipped two open-source Blockly accessibility plugins enabling independent programming for blind and low-vision students.
  • Led an IRB-approved iterative co-design study across two phases; participants progressed from unable to use the platform to completing core tasks without assistance.
  • Standardized pilot study data across projects into a reusable dataset schema; automated data logging with Python scripts.

03 โ€” PROJECTS

Projects

AI / ML

Exploring Practical Applications for LLM-Generated Synthetic Data

Designed and evaluated a reusable workflow for generating synthetic data with LLMs across two use cases: tabular augmentation and human-behavior simulation.

Fairness / Ethics

Ethical AI in College Essay Scoring

Built an AI-powered essay evaluation tool using GPT and Gemini, then asked the harder question: where does it fail?

Education / ML

Modeling Course Success with Probabilistic Programs

Built a Bayesian generative model predicting student success from study behaviors.

Computer Vision

Computer-Aided Diagnosis: Diabetic Retinopathy Detection

Developed a ResNet50-ViT hybrid model achieving 97% accuracy on the APTOS 2019 benchmark dataset.

Computer Vision

The Effect of Limited Early Color Perception on Object Detection

Comparing human and machine object detection performance when color information is restricted during early training.

Education Research

Student Voices: Surveying Principles for a New Educational Institution

A comparative survey study across MIT, Harvard, and Roxbury Community College examining student perspectives on proposed principles for an affordable new educational institution.

04 โ€” TEACHING EXPERIENCE

Teaching Experience

Inspirit AIIndustry

Instructor โ†’ Weekend Lead, Program & Curriculum Manager

Mar 2021 โ€“ Present
  • Taught foundational AI/ML concepts to 320+ students across 32+ cohorts; mentored student projects spanning object/cancer detection, distracted-driver analysis, and voice interfaces, guiding the full ML pipeline from data preparation through responsible-AI evaluation.
  • Maintained and redesigned portions of the AI/ML curriculum; reviewed and debugged course materials and prepared instructor-facing resources, consistently receiving highly positive feedback.
  • Progressed from Instructor to Program Manager, Curriculum Manager, and ultimately Weekend Lead; managed cohorts of 250+ students and led scheduling, onboarding, and coordination efforts.
  • Provided personalized 1:1 instruction, adapting AI/ML concepts to each learner's background and pace.
MITCourse Staff

Lab Assistant & Grader

6.00Introduction to Computer Science and ProgrammingLab Assistant
9.40Introduction to Neural ComputationLab Assistant & Grader
6.C01Modeling with Machine LearningGrader
IEEE-HKNHonor Society

Member & Peer Tutor ยท Eta Kappa Nu

2021 โ€“ 2022 ยท MIT

Selected as a member of the MIT chapter of HKN, the electrical engineering and computer science honor society. Provided peer tutoring to fellow members across a variety of EECS courses, adapting explanations to each student's background and pace.

06 โ€” CONTACT

Let's talk.

I'm always open to research collaborations, conversations about my work, or just connecting with people doing interesting things. I'd love to hear from you.

GitHubLinkedInCV / Resume

ยฉ 2025 Samar Abu Hegly