Agentic AI Alignment
Role: Conversational Designer
Problem: ServiceNow’s enterprise generative AI platform faced critical deployment barriers: multi-turn context drift, ungrounded responses, and susceptibility to adversarial prompt injections. Simultaneously, an escalating taxonomy bottlenecked the Human-in-the-Loop (HITL) data curation workflow, stalling feature validation and product time-to-market. The challenge required designing guardrails for user trust while optimizing internal model training pipelines.
Process: Developed gold-standard datasets, engineered adversarial guardrails, and aligned engineering teams around data taxonomies.
Outcome: Calibrated HITL evaluation UI, driving a 25% increase in cross-functional production throughput.
Methods & Tools: Jupiter Notebooks, ServiceNow Environment, Snowflake, Sreamlit, Prompt Injection, Empircal Testing, Persona Development, Taxonomy Creation, Adversarial Testing, Conversational Design, Python, YAML, JSON, Python.
Client
ServiceNow
Year
2024
ETL Data Pipeline Engineering
Problem: Unstructured airport accessibility PDFs blocked automated data mapping for user dashboards.
Process: Orchestrated HITL dataset alignment and engineered FAISS-backed RAG architectures for deterministic retrieval.
Outcome: Automated 100% of manual parsing, delivering high-fidelity datasets for downstream visualization.
Methods & Tools:
Client
The Atlas Project
Year
01/01/0001
ADA Title II Compliance
Role: Web Content Design Strategist
Problem: Disjointed accessibility standards and unstructured data bottlenecked cross-functional frontend UX deployments.
Process: Calibrated WCAG audits, orchestrated component libraries, and aligned UX/engineering via JSON sitemaps. • Outcome: Mapped 1,300+ enterprise pages into JSON, accelerating targeted frontend remediation and governance.
Methods & Tools: User Interviews, Requirements Gathering, Drupal, SiteImprove, Google Anaytics, Adobe Creative Cloud, HTML, CSS, WAI-ARIA.
Client/Company
Estrella Mountain Community College
Year
2025
UX Writing for AI
Problem: Google Gemini required massive volumes of highly constrained, gold-standard human-alignment data to mitigate model regressions, safety failure modes, and edge-case behaviors in high-risk query categories.
Process: Engineered thousands of precise training data points, stress-tested model outputs for vulnerabilities, and formulated global documentation to scale ethical AI standards across international teams.
Outcome: Expanded Gemini’s safety guardrails and accelerated data pipeline training velocity by standardizing alignment and annotation workflows across global curation teams.
Client/Company
Cognizant, Google
Year
2023
Infotainment System Usability Testing
Explore a curated collection of our past work, where imagination meets strategy. Each project reflects our drive to deliver thoughtful, effective solutions.
Client
Hyndai
Year
2018
Micro-Learning Usability Research
Explore a curated collection of our past work, where imagination meets strategy. Each project reflects our drive to deliver thoughtful, effective solutions.
Client
U.S. Dept. of Defense
Year
2021
Rehabilitation App and Electronic Health Record
Role: User Experience Research Intern
Problem: Remote stroke rehabilitation apps and wearables lacked a cohesive user architecture that balanced complex clinical requirements with strict HIPAA and WCAG digital accessibility compliance.
Process: Map holistic user journeys, lead cross-functional design thinking sessions, and conduct iterative usability studies combining qualitative survivor interviews with quantitative occupational therapist surveys.
Outcome: Aligned product roadmaps with clinical needs and delivered validated UI prototypes that guaranteed absolute data privacy and digital accessibility compliance.
Client/Company
Healing Hand Tech
Year
2022-2023
Protoype of Mobile Chat Interface
ASU requested that II submit a working prototype of a chat application for a UX role they were hiring for.
Client/Company
ASU
Year
2017
Problem: Developing reliable spatial interactions for the MetaQuest required eliminating sensor noise, data outliers, and device connectivity failures that corrupt multimodal research datasets.
Process: Orchestrated hardware setups—including motion capture, eye trackers, and headsets—across 200+ high-fidelity research sessions while troubleshooting live Python scripts and OS compatibility roadblocks.
Outcome: Delivered clean, bias-mitigated spatial interaction datasets that optimized hardware performance, maintained rigorous quality assurance standards, and secured data transfer pipelines.
Usability Testing for Meta Quest VR
Client
Experis, Meta
Year
2022
Social Media Story Design
Adobe Illustrator, Photoshop, and Adobe XD.
Client
None
Year
2018
Ideation and Design
Pencil sketch and scalable vector graphic in Adobe illustrator.