I turn ambiguity into interfaces.
I am an AI-native UX designer. I take complex product ideas and build working, testable prototypes — in days, not months.
Let's talk →Currently taking on select projects · 2026
PROCESS
Two to four weeks. One prototype. Every question surfaced.
Most design engagements start with weeks of discovery, continue through months of iteration, and end with a Figma file someone still has to build. That model made sense before AI changed the math.
I work in focused sprints. You bring the product idea and whatever research you have — user interviews, competitive analysis, napkin sketches, a pitch deck, or just a strong thesis. I bring fifteen years of UX strategy and an AI-accelerated design workflow that produces high-fidelity, clickable prototypes at a pace that didn't exist two years ago.
The output isn't a presentation. It's a working prototype your team can react to, your investors can experience, and your engineers can build from — along with a testing roadmap so you know exactly how to validate it.
Week 1 · Immersion
Understanding the problem, the users, and the constraints.
Mapping the product architecture.
Week 2 · Build
Rapid prototyping with AI-assisted tools.
High-fidelity screens, real interactions, testable flows.
Weeks 3–4 · Refine + Deliver
Iteration with your team.
Delivery of a complete prototype, a component design system, and a testing roadmap for your post-engagement validation.
The result: a prototype that tells you what to build next — and a plan to prove it.
SELECTED WORK
Three problems. Three prototypes.
01 · MOBILE · WELLNESS
Grove — Relational Wellness
A meditation and habit app built on a single insight: your daily practices form a living system, not a checklist. The founders had eight weeks of behavioral research and a strong thesis. They had no screens. They needed an investor-ready prototype in three weeks.
Client: Mira Wellness · Seed stage · 19-day sprint
Your habits aren’t a checklist. They’re a network — and the connections matter more than the individual items.
Your Habit Web
5 of 7 days active
Your wellness ecosystem \u2014 6 weeks of patterns
Click or tap nodes to explore connections
Streaks create anxiety. Research proves it. Grove replaces the streak counter with a warmth indicator that dims gently instead of resetting to zero.
You've practiced 5 of the last 7 days
Your longest rhythm this month: 9 days
You missed one day after a strong week
Warmth glows softly
6 of 7 days — your rhythm is still warm
Streak resets to 0
340 days → gone. Start over.
You took a 3-day break for a family trip
Warmth dims gently
Your practices are here when you’re ready
Streak: 0 days
“You lost your streak!” — actual notification
You’re having a hard week and only managed twice
A faint glow remains
Even a little practice keeps your rhythm alive
Streak: 2 🔥
Implying those 2 days barely count
Warmth preserves momentum. Streaks punish life.
The app knows you slept 6.2 hours. It recommends a grounding session, not a generic meditation. Intelligence felt, not labeled.
Tuesday, 7:42 AM
You slept 6.2 hours last night. This session focuses on restoring calm energy.
12-min Body Scan
Morning Grounding
A full-body scan with progressive attention to restore calm energy.
Breathing Reset
5 min
Quick reset
Focus Prep
20 min
Guided meditation
Drag the slider to adjust session duration
Pattern recognition as self-knowledge: "This was your most consistent week of journaling since November."
WEEKLY REVIEW
Your week in review — Jan 13–19
0
Meditations
0
Walks
0
Journal entries
This was your most consistent week of journaling since November. On days you journaled, you rated your evening calm 2.1 points higher.
What made journaling feel easier this week?
Tap to reflect...
Tap day orbs to see activity details
The screen a user sees after four days away. No guilt. No broken streak. Just: "Hey — no rush."
Hey — no rush.
You've been away for 4 days. Your practices are here whenever you're ready.
Or just browse your patterns →
3-Minute Breathing
Breathe in...
0:00 / 3:00
Even 3 minutes shifts something.
See you whenever.
Tap “Start with 3 minutes” to see the session flow
The complete Grove experience — from first open to weekly reflection — designed and prototyped in 19 days.
Tuesday, Jan 14
Good morning, Mia
5 of 7 days this week
Your warmth is strong
TODAY'S RHYTHM
Habit Web
View →
WEEKLY INSIGHT
“Morning meditation + journaling together: your calm scores are 35% higher on days you do both.”
Week 6 · 4 of 5 days active
How are you feeling today?
Tap to reflect...
The hardest design decision wasn't what to include. It was what to resist. Every wellness app uses streaks. The most important thing I designed was their absence.
19 days · concept to prototype
34 screens designed
Delivered with component design system + testing roadmap
Prototype incorporated into Mira’s Series A investor deck
02 · DESKTOP · FINTECH
Vault — Financial Narrative for Founders
A financial dashboard for startup founders that does something no other tool does: it translates raw data into readable prose. Burn rate, runway, scenario modeling — rendered as sentences a founder can forward to their co-founder, their board, or their investors without spending a Sunday afternoon writing it themselves.
Client: Ledger & Co · Seed stage · 26-day sprint
Not a chart with a tooltip — a paragraph. "You have 14.2 months of runway." Hover any metric to see its source data, trend, and what’s driving the number.
Runway
Now
Start Fundraising
Cash-out
Feb '26
Sep '26
Apr '27
You have 14.2 months of runway at current burn. Your net burn rate is $192K/month, with $48K in monthly recurring revenue offsetting gross expenses.
At this pace, you should begin fundraising conversations by September 2026 to maintain negotiating leverage — assuming a 4–6 month raise cycle.
Recommendation: Maintain current hiring pace. Your runway comfortably supports 2–3 additional hires before the fundraise window narrows.
Cash
$3.21M
Burn
$192K/mo
MoM Revenue
$48K▲6%
Updated today at 2:14 PM · Based on Mercury + Ramp data
Hover highlighted metrics to see trend details
"What if we hire three engineers in April?" Drag the slider. Watch the chart redraw and the narrative rewrite itself — financial modeling as a conversation, not a spreadsheet.
Scenario
What if we add 3 engineers starting April 2026?
Adding 3 engineers at $50K/month increases burn to $242K/month. Runway decreases from 22.3 to 16.6 months. Begin fundraising by Jan 2027 — comfortable with your current trajectory.
Drag the slider to model scenarios · Hover the chart to compare trajectories
One click generates a board-ready memo from raw transaction data. Switch between tones — formal, conversational, or brief — and watch it rewrite in real time.
Board Update
DraftCanopy Labs — February 2026 Board Update
Canopy closed February with $3.21M in cash and a burn rate of $192K/month, providing 14.2 months of runway. MoM revenue grew 6% to $48K, driven by expansion contracts and improved retention across the enterprise segment.
Headcount is 22, up from 19 in January following three engineering hires. The team shipped the v2.0 API migration on schedule. Customer acquisition cost decreased 18% to $1,240 per account.
Strategic outlook remains strong. Current runway comfortably supports the planned Q2 hiring wave. We recommend beginning fundraise preparation in Q3 2026 to maintain negotiating leverage.
Recommendation: Begin Series A preparation in July. Current metrics position us well for a $12–15M raise at favorable terms.
| Metric | Feb | MoM |
|---|---|---|
| Cash | $3.21M | −$192K4.2% decrease from Jan |
| Revenue | $48K | +6%3 new enterprise contracts |
| Burn | $192K | +8%3 eng hires added $49.5K |
| Runway | 14.2 mo | −0.3Still above 12-mo threshold |
Generated in 4 seconds from 847 transactions · Mercury + Ramp + Gusto
Click tone buttons to see the narrative rewrite · Hover table rows for detail
Every number in Vault traces back to a real transaction. This is the live feed — 847 transactions reconciled across four sources. Trust built through radical transparency.
Live Activity
ACH deposit — Acme Corp
+$12,400AWS infrastructure
−$4,280Wire transfer — Payroll
−$67,200Figma — Team plan
−$75Payroll taxes — Feb
−$18,340ACH deposit — Beacon Health
+$8,600Vercel — Pro plan
−$320Monthly close — Jan reconciled
—Stripe payout
+$3,240Travel — SFO → NYC
−$1,240Benefits — Health insurance
−$12,800ACH deposit — Northstar AI
+$22,000847 transactions · $3.21M reconciled
All verified ✓
Hover source pills to filter · Hover transactions to see categories
Six months of cash flow as a stream graph, annotated with the moments that mattered. Revenue is accelerating — February inflows hit 4× September.
Cash Flow — 6 Months
Inflows growing +300% over period. Burn stable at ~$168K/mo.
Total In
$162K
Total Out
$1010K
Net
−$848K
Revenue is accelerating — February inflows hit $48K, 4× September. Burn is rising with headcount but the ratio is improving. At current trajectory, cash flow positive by Q4 2027.
September 2025 – February 2026 · Mercury + Ramp data
Hover months to see inflow/outflow detail · Hover annotations for context
The complete Vault dashboard — narrative overview, scenario modeling, and financial storytelling — designed for founders who need to communicate their numbers, not just see them.
app.vault.finance
Vault
Canopy Labs
Cash
$3.21M
−$192K MoM
MRR
$48K
+6% MoM
Burn
$192K
+8% MoM
Runway
14.2 mo
−0.3 MoM
Sources
February is tracking well. You have 14.2 months of runway. Burn is elevated from January's hiring, but revenue growth is healthy and runway remains comfortable. Begin fundraising by September 2026.
Revenue
$48K
↑ 6% MoM
3 new enterprise contracts · NRR at 114%
Net Burn
$192K
22 headcount
+$49.5K from 3 eng hires · Non-payroll flat
Cash Flow
Feb inflows hit $48K — 4× September. Cash flow positive by Q4 2027 at current trajectory.
The biggest design risk was the serif typeface. Everyone said you can't use a serif in a SaaS dashboard. But Vault isn't a dashboard — it's a document that lives in the browser. The serif was the decision that made it feel different from everything else.
26 days · concept to prototype
28 screens designed
Delivered with component design system + testing roadmap
Prototype became centerpiece of Ledger’s YC W26 application
03 · MOBILE + DESKTOP · EDTECH
Catalyst — AI-Native Learning
An AI-powered learning platform exploring how higher education could work if the tools were designed for how students actually learn — and how professors actually teach. Two complete interfaces sharing one design system, with AI grounded in course materials rather than the open internet.
Client: Lodestar Education · Seed stage · 33-day sprint
A conversational learning interface grounded entirely in course materials. Every response cites its source — visible provenance that builds trust instead of eroding it.
Study Partner
Grounded in Dr. Reeves's course materials
Can you explain how working memory is different from short-term memory?
Great question — Dr. Reeves makes an important distinction here. Short-term memory is a passive store — it holds information briefly. Working memory includes short-term storage but adds active processing.
In Baddeley's model, the central executive coordinates information across subsystems. Dr. Reeves puts it this way: “Short-term memory is where information waits. Working memory is where thinking happens.”
Ask about this week's material...
Tap source chips to see citations · Tap follow-up prompts to continue the conversation
AI-generated exercises that calibrate to each student’s understanding. Every student gets a unique experience — a richer signal of engagement than any tool that came before.
Active Proof
Memory Systems · Exercise 2 of 4
Short Answer
Explain the key difference between the phonological loop and the visuospatial sketchpad in Baddeley's working memory model. Use a specific example for each.
The phonological loop handles verbal and acoustic information — like when you repeat a phone number in your head to remember it. The visuospatial sketchpad handles visual and spatial info — like when you...
Previous: Exercise 1 — Reviewed
Good distinction between sensory and short-term memory. Your example of the cocktail party effect was accurate — see Lecture 6 at 12:30 for Dr. Reeves's additional context.
Understanding:
Getting there
2 of 4 exercises
This week: 2/3 threshold
Tap “Check my understanding” for AI feedback · Tap “Hint” for a clue
Where is the class struggling? Which concepts need revisiting? Engagement visualized as a living map — not a spreadsheet of grades.
How your students are doing
Module 3: Memory Systems · Feb 3–14 · 31 students
Questions
Exercises
Time spent
Flagged
Sensory memory
Lecture 5
Short-term memory
Lecture 6
Working memory (Baddeley)
Lecture 7 + Ch. 5
Episodic vs. semantic
Lecture 6 · 18:30
Encoding strategies
Ch. 7
Retrieval cues
Lecture 8
23 of your 31 students have been asking the study partner about episodic vs. semantic memory, and exercise accuracy is running below average on this topic. It comes up in your Lecture 6 around the 18-minute mark.
A worked example comparing episodic and semantic memory in everyday life could help bridge the gap.
Hover heatmap cells for detail · Switch between Understanding and Engagement tabs
COGN 201
Week 6 of 15 · Memory Systems
Ask about this week's material...
Grounded in Dr. Reeves's course materials
This Week
Lecture 8: Working Memory & Attention
52 min · Not watched
Reading: Baddeley & Hitch, Ch. 5
~35 min · In progress
Active Proof
Memory Systems
2 of 4
Encoding
Not started
This week: 2 of 3 exercises · 1 of 1 conversation
catalyst.edu/dashboard
Catalyst
Good morning, Dr. Reeves
Module 3: Memory Systems · Week 6 of 15 · 27 of your 31 students active this week
Weekly Summary
Your class is doing well with working memory concepts. The study partner is getting good use — conversations up 12% this week. One area to watch: episodic vs. semantic memory is generating a lot of questions and exercise accuracy is below average.
How your students are understanding each concept
1 concept may need support
Students who might need a check-in
27 of 31 on track
Jordan M.
Hasn’t started exercises this week
Riley T.
Flagged 2 AI responses as confusing
Sam K.
Active in chat but struggling on exercises
What's happening in your class
Alex is exploring encoding strategies with the study partner
6 min ago
23 students have questions about episodic vs. semantic memory
trending this week
2 AI responses could use your review
Sarah completed all 4 exercises for working memory
1 hour ago
New discussion thread on Baddeley’s model has 8 replies
2 hours ago
Two interfaces, one system. The student experience and professor workspace — designed together because learning doesn't happen on one side of the screen.
Fifteen years of building tools for higher education taught me what's broken. This is what I'd build if I could start from zero — a platform where the AI serves the course, the professor stays in control, and the student experience is finally designed for students.
33 days · concept to dual-platform prototype
48 screens designed · 26 student + 22 professor
Dual interface · mobile-first student + desktop-first professor
Active Proof identified as the product’s defining feature
ABOUT
I started as a writer. I still am.
I got my graduate degree in English literature. My thesis project connected the work of Bertolt Brecht with Deleuzian rhizomatic theory…which is an awfully pretentious way of saying I've always been a little obsessed with the ways in which complex systems interact with their intended audience to create meaning. That instinct never left. It just found a different medium.
I spent fifteen years building a UX culture within IT at a prominent higher education institution. Among other efforts, I led the design and development of an open-source learning management system used by millions of students globally. I learned how to design complex internal- and external-facing systems for resistant organizations, how to build the case for UX when nobody's asking for it, and how to ship work that actually matters within institutions that move slowly.
Now I work at the intersection of design strategy and AI-accelerated prototyping in a startup environment. The tools I work with have changed dramatically in the last two years — AI has compressed what used to take months into weeks. What hasn't changed? The grounding instinct that caused me to obsess over critical theory back in school. I still start with the same framing question: what does the person using this actually need? The difference now is that I can answer that question with a working prototype in weeks, not months.
Good design is good writing. Both are acts of clarity.
Let's build something.
I take on a small number of sprint-based design engagements — typically two to four weeks, focused on getting from product idea to testable prototype. If you're building something complex and need it to feel clear, I'd like to hear about it.
kyle@kyleblythe.com →Based in [city] · Available for remote engagements worldwide