Honest field notes from the UX/UI Certificate program at Pierce College. What I learned about AI, automation, design tools, and the difference between hype and practice — recorded as discoveries, not as marketing.
A portfolio shows the work. A discovery journal shows the thinking. This document records what I actually figured out during the program: what AI is, what automation is (they are not the same), what is expensive, what is cheap, what is real, and what is hype. The goal is to leave the program with clarity I can take into a working career, not just a stack of completed assignments.
Verified milestones from my Pierce College academic record.
| Milestone | Date | Detail |
|---|---|---|
| First Pierce enrollment | 2018 | Began coursework toward Associate in Applied Science · Digital Design |
| Letter of recommendation from Min Pak | January 8, 2022 | Pak wrote to Pierce County Economic Development. Described me as "naturally talented and hardworking" with "an incredibly astute attention to detail." |
| AAS Digital Design conferred | June 17, 2022 | Honor Graduate · Chancellor's List · Phi Theta Kappa Honor Society. 4.0 GPA within Digital Design coursework. 3.640 cumulative. |
| Return to school | August 2024 | Enrolled in the UX/UI Certificate to add product design and design systems thinking to a decade of physical-design experience. |
| Current coursework | 2025 – 2026 | DDSGN 215 Advanced UX Design · DDSGN 230 Portfolio. Graduation planned June 2026. |
Approximately 190+ college credits earned across this program and prior coursework. Each DDSGN course is typically 5 quarter credits, representing 50+ hours of contact time and 100+ hours of independent practice.
Direct quotes from Professor Pak's documented feedback. These are the lines I keep going back to.
"Critical thinking is your strongest skill. Build custom design systems to stand out from generic UI components. Strong case studies are the most important strategy."
Min Pak, M.F.A. · DDSGN 215 feedback to me, 2026
"Mila has a very methodical creative process. She develops her concepts thoughtfully, creates interesting exploratory designs, and builds meticulous concepts before she commits herself with confidence."
"Her skills of craftsmanship are highly developed, and she has an incredibly astute attention to detail. Mila creates designs that are cohesive in form and eloquent in their communication of sophisticated concepts."
"Mila is an incredibly well-rounded design student who has found a way to balance all the aspects of the creative process with apparent ease."
Min Pak, M.F.A. · Letter to Pierce County Economic Development, January 8, 2022
Additional Pak feedback that shaped specific decisions in this portfolio will be recorded here as I add the email-by-email log. For now, these documented quotes are the spine.
Most of the confusion in the industry right now comes from people using these two words as if they mean the same thing. They do not. This is the cleanest distinction I learned this term.
A computer following a fixed set of rules someone wrote. Same input always produces the same output. Predictable, testable, cheap to run.
Examples in my workflow.
Cost. Effectively free. Pay once for the script, run forever.
A computer making a probabilistic guess based on patterns it learned from massive amounts of data. Same input can produce different outputs. Often useful, sometimes wrong, expensive to run at scale.
Examples in my workflow.
Cost. Subscriptions, API costs, or compute time. Real money per use.
The shorthand I keep: automation does what I told it to. AI does what it guesses I want. Both have a place. Confusing them is where bad design decisions come from.
Honest cost framing from the inside of the program. Prices change, but the order of magnitude does not. The pattern that matters: software is cheap, time is expensive, bad judgment is the most expensive thing of all.
| Category | What it costs | What it buys |
|---|---|---|
| Figma · Education Plan | Free with school email verification | Full Professional plan: unlimited files, version history, advanced prototyping, libraries |
| Adobe Creative Cloud · Student | ~$20/month student rate | Photoshop, Illustrator, InDesign, Acrobat, Dreamweaver, Firefly, Lightroom, XD |
| Bootstrap framework | Free, open source | Complete responsive grid system, components, JavaScript behaviors |
| ChatGPT Plus | ~$20/month | Image generation, drafting copy, ideation |
| Claude Pro | ~$20/month | Code pair-programming, long-document review, design system debugging |
| Pierce College tuition · WA resident | Roughly $1,200 per quarter (part-time, depending on credit load) | Real instructor feedback. Real rubrics. Real degree. |
| Time | 50+ contact hours per course · 100+ independent | The thing money cannot replace. The reason the work is real. |
| Bad judgment | The cost of doing the wrong thing for six weeks before catching it | Nothing. That is the price tag. This is why design systems exist. |
The cheapest tool in the stack is the one that lets me catch a mistake before it compounds. The most expensive thing is shipping a system that looks fine on the homepage and falls apart on page nine.
Recorded from real conversations I had this term, at school and at work. Each misconception is the long-form version of a single sentence I kept hearing.
"AI can replace designers."
AI can replace certain tasks a designer does. Not the designer. The taste, the systems thinking, the brand judgment, and the user empathy are still the human's job. AI is a microphone, not a voice.
"AI is free."
The free tier of any AI tool is a marketing decision, not a real cost picture. Once you use it daily, you pay $20 to $100 per month per tool. And the bigger cost is the time spent prompting, judging, and editing what it returns.
"AI is always faster."
AI is faster when the task is high-volume and reversible. For one-off decisions with downstream impact, AI is often slower because you spend the time prompting, evaluating, re-prompting. A clear hand sketch is sometimes the fastest tool in the room.
"AI understands your brand."
AI understands what it has seen in training data. Your brand was not in the training data. AI will default to whatever generic pattern looks closest to your brand and call it a match. The designer holds the brand. Every time.
"You can tell when something was AI-generated."
Sometimes you can, and the tells are visual: purple gradients, em-dashes, glowing hovers, generic stock-photo composites, "Sarah Chen" testimonials. But a designer who edits the AI output thoughtfully can make it indistinguishable. The tell is not the technology. The tell is the lack of editing.
"AI does not need a design system."
The opposite. AI without a design system produces drift. Every prompt returns a slightly different palette, type pairing, spacing rhythm. A designer with a strict system can use AI cleanly because the system catches and corrects the drift. A designer without a system gets eaten by it.
Honest per-tool reflection. What I came in believing, what I discovered, and how I use each tool now.
What I learned. Figma is the industry standard because it is collaborative-first. The free Education Pro plan is a real benefit, not a discount.
What I use it for. Wireframes, hi-fi screens, component libraries, Auto Layout, dev mode handoff, brand boards.
Discovery. Auto Layout is the single most important feature. It turns a static screen into a responsive system before you even export.
What I learned. Dreamweaver gets dismissed as "old" but it is the cleanest path from code-view to SFTP-deploy in one window. For school portfolios it is the right tool.
What I use it for. Writing HTML and CSS by hand, managing site definitions, SFTP transfer to Pierce server.
Discovery. The Site Definition + SFTP integration is what most students underuse. It saves an entire FTP-client install.
What I learned. Photoshop is still the right tool for any pixel-level retouch, but for layout work, Figma is faster.
What I use it for. Photo retouching, background cleanup, image compression for web, masking.
Discovery. Generative Fill is genuinely useful for product mockups. Generative Background Replace is hit-or-miss.
What I learned. Vector logic is different from raster logic. Illustrator forces you to think in shapes and paths, which makes you a better designer.
What I use it for. Logo work, icon design, vector illustration, print layouts that need infinite scale.
Discovery. Building an icon system in Illustrator and exporting as SVG (not PNG) is what makes a portfolio feel crisp.
What I learned. Excellent at first-pass exploration. Mediocre at staying inside a brand. Requires editing.
What I use it for. Hero image variations, brainstorming, draft copy.
Discovery. The image generation drift is real. After 4 to 5 iterations, results blur into generic. Use early in the process, not late.
What I learned. Strongest in code and long-form reasoning. Excellent at catching my own bad habits in CSS and copy.
What I use it for. CSS architecture, design system audits, debugging, copy rewriting for tone consistency.
Discovery. Treating Claude as a pair-programmer (I propose, it pushes back) produces better work than asking it to just generate.
What I learned. Tightly integrated with Photoshop and Illustrator. Better at staying on-brand than ChatGPT image gen because you can guide it with reference images.
What I use it for. Background expansion, texture generation, material moodboards.
Discovery. Reference-image-guided generation is the part most students skip. It is where Firefly beats other tools.
What I learned. Bootstrap is not the design. Bootstrap is the scaffolding. The design happens in the CSS you write on top of it.
What I use it for. 12-column grid, responsive breakpoints (col-md, col-xl), carousel, cards, jumbotron, navbar — Pak's rubric components.
Discovery. The rubric components do not have to look like the documentation examples. Bootstrap is meant to be overridden.
What I learned. Small Python scripts are the cheapest design tool I have. They handle the volume work no GUI can do efficiently.
What I use it for. Renaming 29 files, compressing a 153 MB PDF to 14 MB, recursive SFTP uploads, em-dash sweeps.
Discovery. A designer who can write 20 lines of Python ships work the average designer cannot. This is automation, not AI.
| The Claim | Verdict | Reality from inside the work |
|---|---|---|
| "AI will design your whole website." | Not realistic. | AI can generate fragments. Stitching them into a coherent system is still a human job. The fragments drift without supervision. |
| "AI can write your case study for you." | Partly realistic. | It can draft. It cannot reflect on your actual experience. The drafts read as generic unless a real practitioner edits them in their voice. |
| "You can become a designer in 3 months with AI." | Not realistic. | AI shortens execution time on specific tasks. It does not shorten the time it takes to develop taste, judgment, or empathy. Those still take years. |
| "AI image generation will replace stock photography." | Already happening. | For hero illustrations and moodboards, yes. For real people, real spaces, real product, hire a photographer. The difference shows. |
| "You can ship a portfolio site in a weekend." | Technically yes. | You can ship a Framer template in a weekend. A hand-coded Bootstrap portfolio with a custom design system and real case studies is months. The first one is borrowed authority. The second is yours. |
| "AI does not understand design principles." | Half true. | It recognizes Gestalt patterns it has seen. It does not apply them with intention. A designer with the principles internalized uses AI better than a designer without them. |
| "Manual design is dead." | Not even close. | Manual sketching, wireframing on paper, and hand-written CSS still produce the cleanest decisions. Manual is slower per artifact, but it is faster per correct decision. |
| "You need a college degree to be a designer." | Not strictly true, but the degree changes what you can demand. | A degree gives you a vocabulary, a rubric, and a documented practice. Without it, you can still design. With it, you can defend your design choices in a job interview. |
Words like Gestalt, progressive disclosure, information architecture, semantic markup, design tokens, Auto Layout, responsive breakpoints. I could already do many of these things instinctively from ten years of flooring work. The program gave me the names. Names let me defend the choices in a meeting.
Before the program I knew when something looked off. I could not always explain why. Now I can: hierarchy is broken, contrast is too low, spacing rhythm is inconsistent, the type pairing is fighting the layout. The rubric is a tool I will use every day for the rest of my career.
Before the program I was using all three modes without thinking about which was which. Now I have a framework that tells me when each mode is appropriate. That clarity is the single biggest jump in productivity I have made.
Five colors, two fonts, three sizes, two radius values, three pixels of hover lift. These constraints were the hardest thing to learn and the most valuable. They are what separate a portfolio that looks designed from one that looks decorated.
Professor Pak does not soften feedback. "Critical thinking is your strongest skill" is a high compliment. It is also a challenge: every design choice has to be defensible by reason, not feeling. The bar she set is the bar I now keep for myself.
Even a one-page mockup gets a five-color palette, a two-font pairing, and a three-size scale defined first. The system is the editor. I am the system.
First drafts, batch operations, exploration — AI handles all of it. Brand decisions, user flow, system architecture — those stay mine. If a decision compounds across the product, a human takes the first pass.
The reasoning is fresh during the work and forgotten after. Case study notes get captured the same week, not the same month.
A polished screen impresses someone for ten seconds. A system that holds across twenty screens impresses someone forever.
What I take into the next phase of my career is not just the certificate. It is the clear distinction between manual, automation, and AI. The discipline of restraint. And a professor's honest feedback that shaped every visible and invisible decision in this body of work.