Itir Sayar

Advisors: Dr. Ravi Karkar · Dr. VP Nguyen

Chancellor's Award Scholarship (merit-based, ~$72k over four years) · Dean's List

Honors Thesis: Lymphoma Survival Prediction — ML/radiomics pipeline for progression free survival prediction in Diffuse Large B-Cell Lymphoma (DLBCL); submitted to SNMMI AI 2023 Challenge

• Built a multimodal deep learning pipeline (En3) for engagement classification during educational video watching; co-first author on manuscript currently under revision.
• Leading a music perception study examining whether EEG and EDA signals can decode musical factors such as genre, tempo, mode, and register.
• Designing and conducting user studies that collect multimodal physiological signals (EEG, EDA, and eye tracking) using consumer grade wearable devices, with custom built experimental software and participant survey instruments.

• Contributed AI data labeling to a child online safety project measuring exploitation indicators in mobile app store reviews at scale; work supported development of app-danger.org.
• Project research was covered by The New York Times and The Wall Street Journal.

• Developed ML/radiomics pipeline for progression-free survival prediction in DLBCL using 3D PET imaging; work formed the basis of the honors thesis and an entry into the SNMMI AI 2023 Challenge.
• Processed 3D PET scans into 2D slices using MATLAB with Gaussian filtering; built deep learning models for image classification and resolution enhancement.

• Contributed to a biologically inspired reinforcement learning approach combining muscle memory with TD3 + dynamic gating control algorithms, reducing computational cost by ~60% while preserving performance across long action sequences.
• Presented work at CICS First Friday Fair (poster).

• COMPSCI 563 — Internet Law & Policy: Managed grading and in class activities.
• COMPSCI 365 — Digital Forensics: Supported syllabus and assignment development.
• COMPSCI 345 — Data Management: Led lab sections; guided 197 students through SQL and database concepts, held office hours.

• COMPSCI 345 — Data Management: Assisted students alongside two professors; led labs, held office hours, authored and proctored exams.

• Leading undergraduate research team on a year-long Simulated Personalized Health Insights project; mentoring scholars in research methodology, physiological data simulation, and academic career development.

• Alex Goldstone (UMass CS Undergraduate, 2024 – 2025) — Honors thesis mentee.
• Patrick Do (UMass CS Master's, 2024 – 2025) — University of Notre Dame CS Ph.D. program (2025).

• Conducted forensic data extraction and evidence flagging for active legal cases using Cellebrite and Magnet Axiom.
• Built a Flask web application for automated legal case report generation, reducing report creation time.
• Led data recovery research on encrypted and deleted files using open-source forensic tools; led onboarding training for new lab interns.
• Volunteered at the 2023 and 2024 National Cybercrime Conference, supporting session logistics and coordination.