Juliana Souza

Welcome — I am

Juliana Souza

PhD Researcher in Human-Computer Interaction & Software Engineering

I study how people resume tasks, recover digital context, and manage artifacts during fragmented knowledge work — and I build tools that reduce the cognitive cost of interruptions.

Download CV Contact me

About Me

I am a PhD researcher working at the intersection of Human-Computer Interaction, Software Engineering, and productivity tools. My research focuses on a challenge that affects almost every knowledge worker: the high cognitive cost of context switching. When people are interrupted — by meetings, messages, or competing tasks — they lose the mental state they had built up, and resuming work takes considerable time and effort.

I investigate how people resume tasks after interruptions, how they recover relevant digital artifacts and context, and how intelligent tools can reduce this friction. My work combines empirical user studies, system design, and machine learning to build and evaluate tools that help people work more effectively in fragmented environments.

I am based at the University of Zurich, where I conduct research that bridges academic rigor and practical impact. I am open to research collaborations, industry partnerships, and opportunities at the intersection of HCI, SE, and intelligent productivity tools.

Empirical & Design Research
Cognitive Science & Productivity
Prototype & Tool Development
Quantitative & Qualitative Studies

What I Do

Core areas of expertise and research practice

Human-Computer Interaction

Designing and evaluating interactive systems with a focus on user needs, cognitive load, and usability in real-world contexts.

Software Engineering Research

Studying developer workflows, tool support, and the socio-technical dimensions of software development at scale.

Task Resumption & Context Recovery

Understanding how people reconstruct interrupted tasks and building tools that capture, store, and surface the right context at the right time.

User Studies & Experimental Design

Running controlled experiments, think-aloud studies, diary studies, and surveys with rigorous methodology and ethical practice.

Data Analysis & Mixed Models

Applying statistical methods — including linear mixed-effects models — to analyze behavioral and performance data from user studies.

Prototype Development

Building functional research prototypes and tools that go beyond mockups — evaluated with real users under realistic conditions.

ML for Artifact Recommendation

Designing machine learning pipelines that infer task context from interaction history and recommend relevant files, documents, and resources.

Academic Writing & Literature Reviews

Producing clear, well-structured research papers and systematic reviews — published and reviewed at top HCI and SE venues.

Education

PhD 2023 – Present

PhD in Informatics

University of Zurich — Zurich, Switzerland

Research focus on task resumption, digital context recovery, and artifact management for knowledge workers. Combining empirical studies with tool design and evaluation.

PhD 2022 – 2023

PhD in Informatics

Virginia Commonwealth University — Richmond, United States

Research focus on focusing on image classification, mixed reality, cognitive psychology, and project-based learning.

MSc 2016 – 2018

Master of Science in Informatics

State University of Maringa — Paraná, Brazil

Research focus on pattern recognition and image processing. Included thesis, 9 courses, and a teaching internship in Algorithms and Data Structure.

BSc 2011 – 2015

Bachelor of Science in Informatics

State University of Maringa — Paraná, Brazil

Included thesis, 54 courses, internship in Android Soft-ware design and development at PhDRisk LTDA.

Experience

Research 2023 – Present

PhD Researcher / Research Assistant

University of Zurich — Zurich, Switzerland

  • Designed and conducted controlled user studies investigating task resumption behavior after interruptions in knowledge work contexts.
  • Built functional research prototypes (TaskSnap, MeetCapsule, Momentum) and evaluated them with real users.
  • Analyzed behavioral data using linear mixed-effects models and qualitative coding.
  • Published and submitted research papers to top-tier HCI and SE venues.

Research Projects

Selected tools and systems I have designed, built, and evaluated

TaskSnap

Published

A lightweight desktop tool that captures a task snapshot — a structured summary of open artifacts, browser tabs, and notes — so that knowledge workers can resume interrupted tasks quickly and with less cognitive effort.

Problem Interruptions cause workers to lose task context, and resumption is slow and error-prone without structured support.
Contribution Designed and implemented TaskSnap; designed and ran a controlled user study (N = [REPLACE]) measuring resumption time, errors, and perceived effort.
Methods Controlled lab experiment, linear mixed-effects models, qualitative coding.
Technologies Electron, TypeScript, Jupyter, SQLite
Paper DOI

MeetCapsule

IN PROGRESS

A tool for recovering meeting-related artifacts and context after collaborative sessions, helping workers reconnect with action items, decisions, and shared materials without manual notetaking overhead.

Problem Meetings generate context that is often lost or scattered; workers struggle to reconnect to pre-meeting tasks afterward.
Contribution Designed the artifact recovery pipeline and the user interface; conducted formative and evaluative studies with knowledge workers.
Methods Diary study, semi-structured interviews, co-design sessions, prototype evaluation
Technologies Electron, TypeScript, Jupyter, SQLite

Momentum

In Progress

A context inference engine that analyzes recent interaction history — file accesses, browser activity, and application usage — to proactively surface relevant artifacts when a worker returns to an interrupted task.

Problem Workers cannot always anticipate which snapshot to capture; Momentum infers the relevant context automatically from behavioral signals.
Contribution Designed the ML pipeline for artifact ranking; evaluated recommendation quality using precision/recall metrics and a user study.
Methods Machine learning, user evaluation, offline dataset analysis
Technologies Electron, TypeScript, Jupyter, SQLite

Publications

Peer-reviewed papers, workshop contributions, and preprints

View all on Google Scholar

Skills

Research Methods

Controlled Experiments Think-Aloud Studies Diary Studies Semi-Structured Interviews Surveys Usability Evaluation Thematic Analysis Grounded Theory

Programming

Python JavaScript TypeScript Java HTML / CSS SQL R

Data Analysis

Linear Mixed-Effects Models Regression Analysis ANOVA Descriptive Statistics R (lme4, tidyverse) Python (pandas, scipy) Qualitative Coding

Tools & Platforms

Git / GitHub VS Code Electron React Figma LaTeX Jupyter

Get in Touch

I am open to research collaborations, academic opportunities, internships, and industry roles. Feel free to reach out.

Email

jsouza@ifi.uzh.ch juulianags@gmail.com

Location

Zurich, Switzerland

GitHub

github.com/JulianaGSouza

LinkedIn

linkedin.com/in/juliana-souza-6673bb139

Google Scholar

Google Scholar Profile