Product Designer
6 months
Avahi Inc. is an AWS Premier Consulting Partner that helps enterprises adopt cloud, data, and AI solutions at scale. Their services span AWS migration, generative AI/ML, data & analytics, and cloud cost optimization, with a focus on security and compliance.
My role involved designing the end-to-end experience of the tool, creating flows for comparing text and image generations, translating technical metrics into clear visuals, and supporting live demos and client-facing conversations.
Clients were evaluating multiple providers at once, but had no clear way to compare them.
Some clients evaluated text outputs; others needed to compare image generations too.
Migrations required confidence in both performance and cost, not another opaque demo.
Avahi and AWS salespeople needed a tool that gave clients neutral, side-by-side comparisons across providers and modalities, so sales reps could build trust and accelerate adoption.
My contribution: I owned the end-to-end evaluation flow, interaction model, prototyping, and design iteration with product, engineering, and sales stakeholders.
The most important questions were simple: How much will this cost? How fast will it run at scale? Those questions became the product’s two core metrics.
To move quickly toward an MVP, I ran unstructured participatory design sessions with stakeholders. We sketched workflows together, aligned priorities in real time, and validated concepts with the people closest to client needs.
first iteration
I unified upload, processing, results, cost, latency, and report export into one consistent flow. This reduced interaction debt and made the comparison logic easier to explain during demos.
The final evaluator enabled enterprise clients and sales reps to compare providers across text and image generation, run single or bulk evaluations, understand cost and latency at a glance, and export boardroom-ready reports.
Observed in prototype evaluation and stakeholder review: sessions became shorter, cost and speed tradeoffs became easier to explain, and the team had a repeatable language for discussing model performance.
This project reinforced my ability to translate technical benchmarks into business-friendly UX, design for quick demos and enterprise workflows, and recognize when simplification matters more than additional UI polish. Next time, I would validate the metric hierarchy even earlier with real sales conversations.