Verdict: Act
| Company | Score | Role | Score | |
|---|---|---|---|---|
| AI centrality | 3 | Total comp | 2 | |
| Stage / equity window | 1 | Scope expandability | 3 | |
| Design influence ceiling | 3 | Craft depth | 3 | |
| Trajectory | 3 | AI exposure quality | 3 | |
| Company total | 10 | Portfolio value | 3 | |
| Role total | 14 |
Stage scores 1 because Midjourney is self-funded with no disclosed employee equity. Cash comp is $285K–$350K. California doesn't require equity disclosure, so silence in the posting isn't proof of silence in the package. Ask in the first substantive conversation.
Narrative fit: Guler describes Midjourney's mission as technology that helps people "perceive what they mean, feel, or imagine before they can articulate it." That is the intent-setting problem your Trust essay opens on and the Agentic Labs demos solve in production. The framework is already published and linkable, which is the rarer half.
Trigger
Banu Guler, installed as CDO ~July 24 through the Co-Star acquisition, posted the Design Lead role roughly five weeks ago. Two IC postings appeared alongside it: Senior Product Designer and Product Designer. Three product-design roles open simultaneously at a company whose careers page still says projects are led by engineers and the founder. This is a function build. Confidence: high.
Window
Guler is ~50 days in. The first-hire window for a new design executive runs roughly 30 to 90 days, which puts her in the middle of it, and the week after Labor Day is when searches that drifted through August start moving again.
In Issue #10 I told you to wait for the Intent-Based Interaction artifact to ship. That was the right call five weeks ago. It isn't now: the posting has aged, two IC roles have joined it, and the cost of arriving late to a first hire is higher than the cost of arriving with a slightly thinner evidence set. Send with what's published.
What this trigger means for the design org
Guler isn't starting from zero. Chris Beiser has been Senior Product Designer since May 2023, and roughly two dozen Co-Star employees came over in the acquisition. What's missing is the organizational layer: leadership, structure, and enough capacity to work across several product surfaces at once — image and video generation, Medical, a standalone consumer app, and projects that aren't named publicly.
The Design Lead's first job is probably not designing. It's deciding how product design operates across those surfaces: the hiring model, the review cadence, how designers sit alongside engineering-led project teams. Reasoning from pattern here, not from a stated mandate. But the posting asks the hire to recruit the team, and the careers page tells you who currently runs projects.
Your angle
Positioning archetype: AI-native product design. The ideal lead is a forward-looking visual artifact, specifically Intent-Based Interaction. If that's published at junochen.com when you send, lead with it. If it isn't, the Trust essay is the pragmatic substitute: it names the same problem at the framework level and routes Guler to thinking she can inspect rather than a case-study deck. Check the site before you write.
Guler's line about helping people perceive what they imagine before they can articulate it describes the first handoff in your essay almost exactly. The TinyFish cold-start work — a blank input box that gave first-time users no way to express what they wanted, redesigned into a progressive intent layer — is that problem in production. Use it as a verbal bridge in conversation. It is not portfolio proof and cannot be linked.
Support with Agentic Labs as inspectable production evidence. Carrier IQ shows all five handoffs in a regulated domain. Brand Pulse shows output review on live AI results. These are the pages that hold up when someone clicks through.
Alibaba is your secondary anchor. Enterprise scale, and more importantly the mandate-creation story: you named a structural gap, built the research case, secured the remit. It answers the question underneath a lead role — can this person build a function, or only design inside one.
Link strategy matters more than usual here. The posting says the company cares "more about what you've made than where you've worked." Guler will open your portfolio before she decides whether to reply. Point her straight at junochen.com/trust-is-the-new-interface in the DM. The homepage routing problem from Issue #10 still applies: if a visitor lands on work that's NDA-protected and can't go anywhere, that's the impression. A direct link sidesteps the problem entirely.
Who to contact
Banu Guler — CDO, Midjourney. LinkedIn.
Public fingerprint: Thin. Her July 24 X post is the only substantial first-person statement in the past twelve months. She framed the mission as making people "more creative, more knowing, more alive in the world," and drew a line between that and labs pursuing human replacement. Her personal site still lists her as Co-Star CEO. No recent LinkedIn posts, talks, podcasts, or writing on design hiring turned up.
Warm path: None verifiable from public sources. No overlap I can see with BCG DV alumni, the Alibaba design network, Dive Club, ADPList, or Designer Fund Bridge. Your logged-in LinkedIn graph may show mutuals that mine doesn't. If it does, name the mutual in the close of the message, not the opening.
Channel: Cold LinkedIn DM to Guler, plus a formal application through Ashby. Run both. The careers page warns that emailed applications get lost, so Ashby is the safety net and the DM is the thing that differentiates you.
Outreach messages
Warm message (only if your graph shows a mutual):
Hi Banu — [mutual connection name] suggested I reach out. I've been working on the problem your July post described: helping people perceive what they imagine before they can articulate it. The framework at junochen.com/trust-is-the-new-interface maps five handoffs where that intent-setting challenge plays out in AI systems, and the Agentic Labs demos on the same site show it shipped. I'd welcome 20 minutes on how it applies across the surfaces you're building.
Cold message:
Subject / opening line: Designing for intent people can't articulate yet
Your framing in July — technology that helps people perceive what they imagine before they can articulate it — is the design problem I've spent the last two years on, first in production at an enterprise agent platform and since then in published work. The Trust essay at junochen.com/trust-is-the-new-interface maps five handoffs where AI systems earn or lose user trust, starting with intent-setting: the moment someone faces a blank surface and the system has to help them find out what they actually want. The Agentic Labs demos on the same site show what that looks like built. I'd welcome 20 minutes on how intent-setting differs across the surfaces you're building.
What not to say:
- Don't say "building a design team from scratch." Beiser has been there since 2023 and the Co-Star designers came with the acquisition. That phrase tells Guler you read a headline and stopped.
- Don't open on Alibaba or BCG DV pedigree. The posting explicitly deprioritizes where you've worked. Leading with employer names contradicts the evaluation culture you're trying to meet.
Cover letter
As Midjourney expands across Medical, a standalone app, and projects you haven't named publicly, each new surface runs into the same problem: helping people discover and express creative intent they can't yet articulate. How intent-setting works differently across contexts where users arrive with very different goals and fluencies — that's what the Design Lead role is actually about.
That's the problem I've been working on. "Trust Is the New Interface" (junochen.com/trust-is-the-new-interface) maps five handoffs where AI systems earn or lose user trust, beginning with intent-setting: the moment a person faces a blank surface and the design has to close the gap between what they imagine and what they can specify. The Agentic Labs demos on the same site, Brand Pulse and Carrier IQ, show those handoffs running against real APIs and live agents with enterprise customers. At Alibaba.com I led design and research for the B2B platform redesign as Head of Design and Research, North America — $50B+ GMV, +20% daily transactions — after naming a structural gap nobody had prioritized, building the research case, and securing the mandate to redesign every major surface.
Most recently as Head of Product at TinyFish, I shipped an enterprise web agent platform from 0 to 1 in three months, including a redesign of the intent-setting layer to solve a cold-start problem where users faced a blank input with no way to express what they needed. I moved into product to build AI-natively from zero. I'm coming back to design because I want to apply that technical depth to one high-stakes creative product rather than a horizontal platform, and because I want to be in the work again.
I'd welcome a conversation about how the intent-setting framework applies to the surfaces you're building. The full portfolio is at junochen.com.
Resume emphasis
Craft depth: 3. AI exposure quality: 3.
- Lead with: Agentic Labs (Brand Pulse, Retail Velocity, Carrier IQ) and the Trust essay. At a company that evaluates on what you've made, these are what get clicked.
- Follow with: Alibaba.com, Head of Design and Research, North America. Enterprise scale, mandate creation, function-building proof.
- TinyFish: One line in the header, past tense. "Most recently, Head of Product at TinyFish (enterprise web agent platform, Series A) — shipped agentic platform 0→1 in 3 months, redesigned intent-setting layer, worked daily with agent traces and governance in production."
- Keep, but lower: Thermo Fisher mySupply (regulated 0→1, as Product Design Director at BCG Digital Ventures) and Allē (dual-surface consumer scale, same title).
- Trim: Red Cross and Equinox+. Neither maps closely enough to a creative-AI mandate to earn space at the top.
The non-obvious thing
Read the Co-Star job listings. They're unsigned, but they're from Guler's company, and they carry evaluation language the Midjourney posting never surfaces: a designer who studies the available systems and materials and then makes something that feels "unexpected and inevitable." Repeated iteration. Code-based prototyping. Willingness to throw work away. "Work > ego."
The Midjourney posting is generic next to that. The Co-Star listings are closer to how Guler actually evaluates, and they tell you prototyping fluency and a visible iteration trail matter more than a polished case-study narrative. If you have code prototypes or interactive demos, surface them. Most of the people applying will send decks.
Warning signs
Midjourney is self-funded with no disclosed employee equity. The cash floor is strong ($285K–$350K), but with no equity component, total comp is capped at a company whose trajectory would otherwise justify real upside.
All three product-design postings list salary and benefits with no mention of stock, options, RSUs, or profit sharing. Levels.fyi doesn't have enough data to confirm whether employees get equity at all.
"Projects are led by engineers and the founder." The careers page says it outright. Guler's arrival may be changing that default, but engineering-led is the starting condition, and the Design Lead will likely have to establish design's place in project leadership rather than inherit it. Use the interview to find out whether Guler has the backing to change it, or whether the posting is describing something she's still negotiating for.
Competitive landscape
Creative-tool designer (Figma, Adobe, Canva, or a creative-AI startup). Strong visual portfolio, deep craft in creative tooling, and actual generative-workflow product experience, which beats you on the most obvious axis. Weaker on system-level thinking, published design thinking, and function-building. Counter: the mandate is organizational ("recruit the product-design team") and spans Medical, a standalone app, and unnamed projects. Cross-domain system work and a track record of building the function are where you separate. Confidence: moderate.
AI-native designer (Anthropic, OpenAI, another lab). Brand pedigree beats yours. Consumer creative product experience and published work generally don't. Counter: the Trust essay is frontier design thinking in public, which this archetype rarely has, because they design inside companies that don't let them publish. Confidence: moderate.
Consumer product designer (Spotify, Instagram, Pinterest). Better polish and consumer craft. Thinner on agent-system depth and the technical fluency Medical and the unnamed projects will demand. Counter: TinyFish gives you practitioner depth in production AI systems, delivered verbally rather than in the portfolio. Confidence: moderate.
Where you're exposed against all three: there is no creative-AI product in your published portfolio. You haven't designed a generative image tool, a creative canvas, or a prompt-to-output workflow. The reframe — a system-design and function-building mandate across surfaces — is supported by the posting language, and it's the right frame to argue. Guler may still weight creative-tool experience anyway. The Trust essay and Agentic Labs narrow the gap but don't close it, and you should go in expecting the question.
Post-outreach signals
Treat Midjourney on AI-native startup timing:
- Reply within 48–72 hours = healthy.
- Silence past 5 business days = the message didn't land, or the search is further along than the posting suggests.
- A reply that engages the intent-setting framing rather than thanking you = the hook worked. Answer within 24 hours and go deeper.
- A recruiter-only reply with a generic screen request = you reached the system, not Guler. Take the screen and work toward her from inside it.
- Follow up once at day 5, anchored to something new — a product update, a post of hers, a shipped feature. One follow-up. After that, move Midjourney to Watch and put the energy elsewhere.
- Midjourney's September Alpha update: The September 3 changelog introduced a lightbox editor accepting plain-language instructions and reference images — concrete product language to reference if you need a follow-up hook after initial outreach.
- Co-Star listings as evaluation proxy: The Co-Star Senior Product Designer posting says the hire will work across Co-Star and "new experiments," which may indicate how Guler plans to allocate designers between her existing company and Midjourney's unnamed projects.
- Midjourney Medical timeline and scope: The Medical project page targets a San Francisco spa opening at end of 2027 with a full-body ultrasound scanner and consumer app — a product surface radically different from image generation that could expand the Design Lead's mandate quickly.
- Suno v6 as creative-AI comparison point: Suno's September 9 launch added partial editing and model-variant selection for precision versus exploration — the kind of granular creator-control problem Midjourney's design team will face, useful as interview preparation for discussing intent-setting across creative tools.