← Back to Portfolio
Selected Works

Numan Ghuman

Quantitative UX Research · AI Enablement · Learning Experience Design

Core Capabilities

Research Focus Areas

Quantitative UX

Usability benchmarks, A/B testing, and inferential statistics (ANOVA, Regression, t-tests) leveraging SPSS, R, and SQL.

AI Evaluation

Benchmarking LLM workflows (Claude, ChatGPT, Gemini) to measure reliability, mitigate friction, and accelerate enterprise adoption.

Learning & Development

Mixed-methods research scaling educational technology and tracking longitudinal learning outcomes across diverse cohorts.

Case Study 1

Deen Detectives

Gamified Application Retention Study t-tests & Regression

Overview

A quantitative usability and engagement study measuring session length, feature adoption, and 30-day retention for a live gamified application.

Responsibilities

  • Designed and fielded targeted survey instruments.
  • Coded qualitative feedback themes into measurable cohorts.
  • Conducted independent samples t-tests (iOS vs. Android) and regression analysis modeling session engagement predictors using SPSS and R.

The Impact

Analysis informed iterative testing cycles that resolved critical onboarding drop-offs.

SUS 87/100

Achieved a "Good" grade usability rating, placing the application well above the 71 industry benchmark.

Case Study 2

Mizan Companion

FinTech UX Benchmark Study ANOVA

Overview

A mixed-methods usability benchmark study conducted across three diverse financial-literacy segments.

Responsibilities

  • Facilitated study sessions measuring task completion, error rate, time-on-task, and SUS scores.
  • Ran one-way ANOVA and paired t-tests to identify critical design gaps across literacy tiers.
  • Identified strong correlations between financial literacy and perceived usability (r=0.91).

The Impact

Findings directly informed a segmented onboarding redesign.

+13 pts

Increase in overall SUS score post-redesign

-73%

Reduction in session error rates

Case Study 3

Enterprise AI Usability

L&D Adoption LLM Benchmarking Longitudinal Data

Overview

A structured mixed-methods research initiative evaluating technology adoption and the learning experience of AI-assisted workflows across 150+ participants.

Responsibilities

  • Benchmarked leading LLMs (ChatGPT, Claude, Copilot) using structured usability-testing protocols.
  • Measured quantitative adoption metrics across diverse cohorts.
  • Built SQL-supported reporting dashboards to track longitudinal engagement and cohort retention.

The Impact

Synthesized findings into stakeholder recommendations that successfully integrated AI into production workflows.

+21%

Improvement in measured operational efficiency

Case Study 4

Experimental Design

IRB Compliance Confound Control SONA Systems

Overview

A rigorous academic experimental study comparing competing theoretical frameworks (Moral Foundations Theory vs. Dyadic Template).

Responsibilities

  • Served on the Institutional Review Board (IRB) assessing protocol safety.
  • Managed the full participant lifecycle via SONA Systems.
  • Maintained strict protocol fidelity across multi-semester data collection to control for confounds.
  • Performed complex statistical analyses (ANOVA, Pearson correlation, hypothesis testing).

The Impact

Ensured high data quality and statistical power for publication-ready samples while maintaining ethical compliance across all active studies.