Suzi Soroczak | UX Research Leader Mixed Methods * AI-Powered Products * Connected Devices * Healthcare
I’m a Principal UX Researcher with 20+ years in the software industry and 15 years specializing in HCI. My work spans Big Tech (Intel, Microsoft, Google, Amazon), federal government (CDC, White House USDS), and health tech startups. I bridge qualitative depth and quantitative rigor to help product teams make confident decisions — whether that’s a national public health platform, an IRS tax tool used by millions, or a real-time glucose coaching app.
I’m currently open to Principal / Staff UX Research roles. Contact: misssuzka at gmail dot com | LinkedIn
PILLAR 1: Mixed Methods Research
I design research programs that combine qualitative and quantitative methods — choosing the right tool for the question, not defaulting to one approach. Methods I use regularly:
Ethnographic field research & semi-structured interviews
- interviews with PH professionals at 64+ CDC jurisdictions
- site visits for IRS Virtual Vita
- home visits for Tencent America group gaming study
Focus groups & moderated usability studies
- simultaneous English/Spanish/Accessibility waves at USDS
- 750% increase in national engagement with PHDS
Surveys & quantitative analysis & panel building
- 500+ person Public Health Professionals panel at CDC
- 10K casual, mid-, and hard-core gamers panel at Tencent America
- community forum NLP/sentiment analysis in R
Card sorting, heuristic reviews, semantic testing for brand identity, service blueprints, customer journey maps
Data storytelling in PowerBI, Tableau, LookerStudio, and Quarto
CDC Public Health Data Strategy: Designed and moderated focus groups with state, tribal, territorial, and local health stakeholders that expanded national reach of CDC’s data strategy by 750% and cut development timelines by 3 weeks. Blueprinted user workflows for the Palantir / One CDC Data Platform migration across 7 OPHDST divisions.
ChildTaxCredit.gov (USDS/White House/IRS/Treasury): Led qual research program (heuristic reviews, parallel usability waves, inclusive research with users with low vision and cognitive disabilities, card sorting via Optimal Workshop) for a site launched in 12 weeks that reached millions of American families. Managing stakeholders across 4 federal agencies/divisions.
View CTC Outreach Research Report | IRS Virtual Vita Report
Contact me at misssuzka at gmail dot com for full study materials and CDC deliverables
PILLAR 2: AI Experience
I research and build AI-powered products — and I know how to study human behavior around AI systems.
Halebee Coach — Women Build AI 2026 Buildathon (Honorable Mention) I designed and vibe-coded Halebee Coach, a real-time glucose coaching companion that connects to Libre and Dexcom cloud accounts to pull live CGM sensor data and deliver personalized coaching suggestions about food, movement, and lifestyle matched to where your glucose is heading right now.
This project gave me hands-on experience with:
- Designing AI response logic for safety-sensitive health context
- Prompting and evaluating LLM outputs for clinical appropriateness
- Rapid prototyping and user testing of conversational AI interfaces
At CDC, I leveraged ChatGPT and the US Web Design System to produce 508-certified UI designs under resource constraints, demonstrating practical AI-assisted design at a federal scale.
At Autodesk, I built interactive product dashboards powered by NLP — transforming community forum data into actionable product sentiment insights for research and product teams.
PILLAR 3: Devices & Sensor Experience
I have a rare combination: peer-reviewed academic research on sensor-coupled displays and current product ownership of a CGM-connected mobile app.
Halebee.app — CGM-Connected Wellness App (Current) As Principal Product Researcher, I conduct user research to understand how people with Type 2 Diabetes make sense of continuous glucose monitor (CGM) data from Libre and Dexcom devices. My research questions include: What mental models do users bring to glucose trend data? What contextual triggers prompt behavior change? How do we design actionable, non-alarming notifications?
Proactive Displays Research — University of Washington (Published) My academic work at the UW Information School focused on proactive displays: sensor-coupled screens that detect nearby people and adapt their content in response. This is the foundational HCI problem that modern ambient AI and smart device interfaces are still solving. Key publications:
- McDonald, McCarthy, Soroczak, et al. (2008). “Proactive Displays: Supporting Awareness in Fluid Social Environments.” ACM Transactions on Computer-Human Interaction, 14(4). [Top-tier peer-reviewed venue]
- McCarthy, McDonald, Soroczak, et al. (2004). “Augmenting the Social Space of an Academic Conference.” CSCW 2004.
Research approach: mixed methods - qualitative observation combined with survey & web data - to understand both individual and group-level impacts of sensor-aware displays in social spaces.
Background & Credentials
Experience: Intel * Microsoft * Amazon * Autodesk * Google* Tencent America * US Digital Service (White House) * CDC * Halebee
Recent Training:
- Building MCPs in Python (Python.org) 2026
- CSPO Certificate - Passionate Product Leadership (Scrum Alliance) 2024
- Designing Strategy (IDEO U) 2024
- Future State Service Blueprinting (Practical by Design) 2024
- Building Dashboards with Quarto (PositConf) 2024
- Storytelling with Data (ELVTR) 2023
- Data Science Specialization (Johns Hopkins / Coursera) 2018
Technical skills: R, JavaScript, Python, HTML/CSS, PowerBI, Tableau, LookerStudio, Quarto, Qualtrics XM, UserTesting.com, Optimal Workshop
Storytelling with Data
Olympic Games Dashboard — Quarto
Community Forum Sentiment Analysis — R + JavaScript
(Autodesk Research Dashboard screenshots available on request)
