Product designer · Human-centered AI

I design AI that people can check, correct and trust.

I'm Priyanshu Nada, a Computer Science senior at San Francisco State. I take products from research and Figma flows all the way to working React prototypes, and lately I've been focused on AI that shows where its answers come from.

See the work Resume (PDF) pnada.jobs@gmail.com
Based in
San Francisco, CA
Open to
Product design, AI/UX and design technologist roles
Graduating
B.S. Computer Science, May 2027

Selected work

Four projects about giving people the right information at the moment they decide.

Each card plays its project's key moment. The case studies show the research, the decisions and what isn't built yet.

01Fintech · Conversational AI

Flow: banking that disappears

AI money tools for Gen Z that show up inside iMessage at the moment someone is about to spend.

End-to-end design · 4 min readRead case study
02AI redesign · Human in the loop

Capstack: funding data founders can trust

A funding directory kept current by an AI agent, with a person approving anything that could cost a founder a deadline.

Research, design, front-end · 4 min readRead case study
03Consumer AI · Explainable suggestions

Closet: an outfit in one glance

An AI wardrobe for busy students that picks outfits from clothes they own and explains every pick.

Concept, interaction, prototype · 4 min readRead case study
04Healthcare AI · Grounded summaries

Groundwork: the patient’s story before the scan

A context assistant for radiologists that links every line to its source record and never suggests a diagnosis.

Research, interaction, prototype · 5 min readRead case study

Tip: press 1 to 4 to jump to a case study.

In progress

What I'm building now.

Three projects on the same idea: AI you can check before you act on it. Each one goes up here with its prototype, evaluation and limits once it's finished.

  • 05Human-approved workflow agentAn AI agent for one hotel operations task. It proposes actions, a person approves them, and it handles failures and undo honestly.Coming soon
  • 06Inclusive multimodal patternsVoice, text and image input as real choices, tested with keyboard and screen-reader users.Coming soon
  • 07Taking the agent to a pilotThe workflow agent with a risk review, cost model and rollout plan, the way it would ship inside a company.Coming soon

About

I came to design through code, and I still build what I design.

Most of my projects start in Figma and end as working software, because that's where the real problems show up: the empty state nobody drew, the score nobody can explain, the error with no way out.

I also work the front desk at SF State's Center for Global Engagement, helping more than 200 international students with applications, housing and onboarding. Watching people get stuck in forms and email threads every week is a big part of why I care about clear systems.

Right now I'm learning to design for AI specifically: answers that show their sources, agents that ask before they act, and interfaces that work without a mouse.

Design
Figma (components, variants, prototyping), interaction design, information architecture, journey maps
Research
Usability testing, heuristic evaluation, personas, surveys in Qualtrics, Maze
Build
React, JavaScript, HTML/CSS, Python, Git
Learning now
LLM APIs, tool calling, evaluation sets, WCAG 2.2
Education
B.S. Computer Science, San Francisco State, May 2027 · Dean's List
Certifications
IxDF Master Classes: Human-Centered Design for AI · AI for Design Systems, 2026
Google UX Design Certificate, 2025

How I work

Four rules I keep coming back to.

  1. Start at the decision.Find the exact moment someone needs help, then design backward from there.
  2. Go where people already are.Put help inside tools people already use before asking them to learn a new one.
  3. Show, then step back.Give the answer and the choice, then get out of the way.
  4. Make it checkable.Show sources and reasons, so people can catch a wrong answer instead of just accepting it.

Process

  1. Research
  2. Flows in Figma
  3. Working prototype
  4. Test with people
  5. Fix and repeat

Contact

Let's talk.

I'm open to product design, AI/UX and design technologist roles and internships. The fastest way to reach me is email.

pnada.jobs@gmail.com LinkedIn GitHub Resume (PDF)