Skip to the work
Open to lead rolesTehran · remote

I design cloud products people can actually operate.

Six years turning dense infrastructure into interfaces engineers trust — and building the teams that keep them that way.

Shipping at parspack since 2023
Ali Mousavi

Currently

Leading design for parspack — the control panel, billing and identity surfaces behind a national cloud platform.

Tehran, Iran

Trusted with the work at

  • parspack
  • Modelic
  • Dehkade Bazi
  • Moj Secure E-Commerce
(02)

Selected work

Products I shaped end to end — the problem, the calls I made, what shipped. The rest are being written up.

Case study in progress
ARAlgorockHow to gain trust with clarityIn progress
Case study in progress
NPSNPS & CSATHow to raise satisfaction in a product's worst seasonIn progress
Case study in progress
NCNotification CenterOne place that serves them allIn progress
Case study in progress
DSDesign SystemOne vocabulary the whole product speaksIn progress
(03)

Leadership

What I actually changed when I stopped being the one pushing every pixel.

70%

of the product on the design system

measured, not assumed — you don't reach that by reminding people

70

fewer support tickets a day

an FAQ before ticket submission — four off each agent's shift

3

designers led

3–4

week hiring loop

task, not puzzle

  • The title changed; the output didn't. The Notification Center, Projects & Tags, and the AI assistant each started as a PRD I wrote — information architecture first, the permission scoping model, channel decisions argued against how messages actually reach people here, and for the assistant an evaluation strategy agreed before anyone built.

    Design arriving after commitment can only make cosmetic moves. Writing the spec puts design in the room where scope is decided, so the ambiguity gets resolved by the person who has to live inside the interface afterwards.

    What a spec from design carriesartifact
    1. 01 Architecture — IA and scoping model, before any screen
    2. 02 Constraints named as constraints, including the local ones
    3. 03 Who solved this well elsewhere, and what we deliberately rejected
    4. 04 How we'll know it worked — written before the build
    5. 05 Rollout in phases, each proving one thing

    Design that writes the spec never argues about sequence.

  • Three designers don't need a weekly ceremony. They need the bar written down, so it holds whether or not I'm in the room.

    Ours is a checklist that runs from before design starts to after delivery. Direction gets validated before anyone goes high-fidelity. Every state and edge case gets designed, not just the happy path. Both languages, RTL checked. Peer review on flow and UI before it reaches me, and a final check with me as the gate. It reads bureaucratic until you notice that seventy percent of the product now sits on the new design system, which is not a number you reach by reminding people.

    Design checklist — the gate every screen passesartifact
    1. Pre-design — PRD with estimation · competitor and pattern research · flows, edge cases, error states mapped
    2. Design — direction validated before high-fidelity · every state designed · system used correctly · desktop and mobile · EN and FA, RTL checked · micro-copy and typos
    3. Review — peer review on flow and UI · handoff doc with naming, states, behaviours · final check with me
    4. Delivery — filed in the shared file · PM notified with name, description, Figma link

    Written down so the bar holds whether or not I'm in the room.

  • I don't hand a designer the solution. I hand them the problem and let them find the way through. It costs a few rounds of back-and-forth at the start, and it's the only thing that actually teaches someone the product rather than teaching them my taste.

    Two notes come up more than any others: we already solved this somewhere else — why are we inventing it again, and, on genuinely new features, this is the first idea, not the best one. Direction isn't mine to hand down either. Everyone puts their read on the table, gut feel included, and we work the trade-offs out loud until one approach is clearly better argued than the others.

    How a direction gets chosenartifact
    1. 01 Everyone states their approach, including the instinct behind it
    2. 02 Trade-offs written down, not defended in conversation
    3. 03 The option we're rejecting is named, with the reason
    4. 04 Whoever owns the work makes the call
    5. 05 Disagree and commit — nobody relitigates it in review

    The designer who found the answer defends it better than the one who was given it.

  • The hard part of leading a small team isn't hiring. It's the moment someone isn't meeting the bar and you have to say so rather than absorb it quietly for another quarter. I've had to make that call and see it through. Three designers means one person below the bar is a third of the team's output.

    The loop that replaced them runs three to four weeks: portfolio review, an online interview, a real task, a review of that task, an HR conversation, then an offer. The task is the part that matters — how someone frames a problem they've never seen tells you more than any case study they've had months to polish.

    Hiring loop — three to four weeksartifact
    1. 01 Portfolio review against stated criteria
    2. 02 Online interview
    3. 03 A real task, not a puzzle
    4. 04 Task review — framing first, craft second
    5. 05 HR interview
    6. 06 Offer

    The case study shows what they shipped. The task shows how they think.

  • Opinions are cheap in a room full of designers. The researcher on my team runs three tracks every week: user research against problems Product raises, analytics and Hotjar on the flows we suspect are leaking, and product research ahead of anything new.

    We measure what we change. Putting an FAQ in front of ticket submission dropped ticket volume three and a half percent — which sounds like rounding error until you run it against two thousand tickets a day and it's seventy fewer, roughly four off each agent's shift. Drop-off gets diagnosed through funnel analysis rather than debated. Pricing patterns went to an A/B test. And the design system sits at seventy percent adoption, measured, not assumed.

    How a decision gets evidenceartifact
    1. 01 Product raises the problem — research plan before any testing
    2. 02 Method chosen to fit the question, not the calendar
    3. 03 Analytics and session recordings for where, funnels for how much
    4. 04 A/B test where the choice is genuinely contested
    5. 05 Baseline taken before ship, compared after

    Seventy fewer tickets a day is four fewer per agent per shift. That's the number that matters.

(04)

The practice

How the work gets made — the parts that survive a change of company.

  • Ambition into a sequence someone can build. I write the PRD, the phased rollout, and the success criteria: what ships first, what each phase proves, what we're willing to be wrong about.

  • Permission scoping, cross-service architecture, operator dashboards. These products fail at the moment of comprehension, so I design for the person under pressure, not the screenshot in the deck.

  • Tokens, component behaviour, Figma-to-code parity, and the RTL and Persian conventions decided once. Seventy percent adopted and measured.

  • Weekly user research, analytics and session recordings, and product research ahead of new features. Research that doesn't change the roadmap wasn't research.

  • Hiring loops, a written quality bar, and mentoring that hands over the problem rather than the answer.

  • I prototype in the real stack and ship interactive mockups, not static screens.

(05)

How I work

Four moves, in order, every time.

01

Frame

Find the real problem and write it down in one sentence everybody signs.

02

Explore

Cheap, wide, ugly. Enough options that the choice is informed rather than first-found.

03

Build

Prototype in the real stack, test with real operators, cut what does not earn its place.

04

Hold

Ship, measure against the baseline we took beforehand, and keep the bar after launch.

(06)

Toolkit

Where the hours went.

Figma

Design, systems, variables

100%

Design systems

Tokens, governance, adoption

95%

Research & testing

Interviews, usability, tickets

90%

Next.js & React

Prototypes that really run

75%

Framer & Webflow

Marketing surfaces, fast

70%

n8n & automation

Ops glue for the team

65%
(07)

Track record

Ten years, from architecture school to a cloud platform's design team.

6

years — designing, most of it in infrastructure

  • Aug 2023 — nowLead Product Designerparspack
  • 2025 — nowDesign ConsultantBoubamzicontract
  • Jul — Aug 2025Product DesignerAlgorockcontract
  • 2024 — 2025Product DesignerModeliccontract
  • 2023 — 2024Product DesignerRamzArzcontract
  • 2023Product DesignerDehkade Bazicontract
  • Feb 2022 — Jun 2023Product DesignerMoj Secure E-Commerce
  • 2020 — 2023Freelance Product DesignerIndependentcontract

95%

client satisfaction

built on trust and results

(08)

Words from clients

People who shipped with me.

“Ali is the first designer here who argued with engineering about cost and won on the merits. The panel rebuild happened because he made it measurable.”
NAN. AhmadiVP Engineering, parspack
“He turned a three-person service desk into a team with actual ownership. Two of the people he hired are now running their own surfaces.”
SRS. RahimiHead of Product, parspack
“Fastest I have ever gone from a fuzzy brief to something clickable. He prototypes in the real stack, so what we discussed on Monday was running by Thursday.”
MKM. KarimiFounder, Cardhub
Open to lead rolesTehran · remote

Let’s make the complicated thing usable.