Static Creative Strategist (AI Image Generation) — DTC Menswear Brand

Please login or register as jobseeker to apply for this job.

TYPE OF WORK

Full Time

SALARY

from 10$/ hour

HOURS PER WEEK

35

DATE UPDATED

Aug 27, 2026

JOB OVERVIEW

Read this first
This is not a "generate images from a prompt" job. We've hired that profile before and it doesn't work for us.

We're looking for someone who thinks like a creative strategist and executes like an art director — someone who can look at a market, understand why an ad works, decide what we should make, and then produce it themselves using AI generation. The strategy and the execution are the same job here.

If you generate good-looking images but can't explain why an ad converts, this role isn't for you. If you understand direct response but your images look obviously AI-generated, it isn't either. We need both.

About us
ARES is a men's underwear brand doing high-volume DTC in the US and Worldwide market. We sell silk-blend boxer-briefs on Shopify and Amazon, and we acquire mainly through Meta Ads, Google, and Applovin. Creative is our number one lever — it decides whether we're profitable or not.

We produce a lot: ad statics, Instagram content, product page imagery. Almost all of it is AI-generated. That's a deliberate production system, not a shortcut. The standard is simple — the final image has to look like a real photoshoot. If someone can tell it's AI, it failed.

What you'll own
Ad statics — you research the market, decide the angle, and build the ad. Layout, hierarchy, copy overlay, visual concept. You bring us ideas, we don't hand you a shot list.
Instagram content — lifestyle and editorial imagery that reads as shot on set.
Product page images — sequential visual storytelling, one narrative beat per image with copy overlay. Not a product gallery.
Realism — skin texture, fabric behavior, lighting coherence, hands, faces, logo accuracy. You retouch and composite until the AI tells are gone.
Consistency — same product, same colorway, same brand codes across hundreds of images. Our product has precise specs that have to be right every single time.
What we need from you
On strategy:

You understand direct response. Hierarchy, focal point, where the eye lands, why one layout converts and another doesn't.
You research competitors and the broader market on your own, and you can defend a creative decision with reasoning instead of taste.
You can take a rough business objective ("we need a back-to-school offer") and come back with a concept, not a question.
On execution:

Deep command of AI image generation — Nano Banana, Midjourney, Flux, Imagen, ComfyUI, whichever you genuinely master. You control the output, you don't roll the dice on it.
Structured, repeatable prompting.
Retouching and compositing (Photoshop or equivalent). Generation alone will not reach our quality bar — the finishing pass is what makes an image usable.
Photographic literacy: lighting setups, lens behavior, depth of field. The people who make AI look real are the ones who understand how a camera actually sees.
Volume without quality drop. We work in batches.
English communication, reliable turnaround, comfortable working from written briefs.
Bonus: apparel or underwear experience, anything where fabric rendering matters. Paid social creative experience.

How to apply
Answer these questions one by one. Applications that send a general pitch instead of answering will not be reviewed.

Ad thinking — Take an ad image you made that performed well. Why did it work? Walk me through what you decided visually and the intent behind each choice. If you have performance data, include it.

Strategy from scratch — You're given one line: "we need a static for a back-to-school offer." No further brief. Walk me through how you get from that sentence to a finished ad — your research, your decisions, your reasoning.

Realism method — What's your actual process for making an AI image undetectable as AI? Name the specific tells you hunt for and how you fix each one, from generation to final file.

Control and consistency — I need the same product, same colorway, same logo, across 100 images. How do you guarantee that? Describe your setup.

Failure case — Describe a project where the AI output kept failing on something specific. What was breaking, what did you try, how did you finally solve it?

Toolstack and day one — Which tools do you use daily and for what? If you started next week, what would your first two weeks look like?

Track record — Which brands or clients have you done this work for? Name them. Send a portfolio with AI-generated work where the AI isn't visible — and if you have before/after (raw generation vs final delivery), send that too. It's what I'm most interested in seeing.

Test:
We don't hire off a portfolio. Shortlisted candidates get a paid test, and we agree the fee before you start.

We're not testing whether you can operate a generator. We're testing whether you can think about an ad and then build it yourself. Read the product section carefully — most of what we reject fails on product accuracy, not on artistic quality.

What to deliver
5 images. Format 4:5. Highest resolution you can produce.
Few brand asset for this test: ----------

Three are briefed. Two are yours.

Image 1 — Outdoor shoot, ancient Greek setting
A man wearing the boxer, an elegant woman beside him. Ancient Greek atmosphere — the brand is named after the Greek god of war, so the reference is ours to own.

How you interpret that is up to you: stone, light, landscape, styling, era treatment, mood. We want to see your taste and your restraint. It should look like an editorial fashion shoot, not a costume set.

Image 2 — Back-to-school sale
A sale creative built around the boxers. Back-to-school timing, US market.

Our offer structure: Buy 2 Get 2 Free ($59.99) · Buy 3 Get 5 Free ($89.99 — $11 a pair).

Two copy rules: never show a percentage discount, and never lead with the single-unit price. Frame the value through the bundle.

Image 3 — Professional studio shoot
A clean studio image. Pro lighting setup, controlled environment.

This is where we judge pure craft — lighting, fabric rendering, product accuracy. Nowhere to hide.

On these three: we want them to look genuinely different from each other. Different lighting, different energy, different treatment. Three variations of the same look tells us nothing. Range is what we're measuring.

Images 4 and 5 — Two Meta ads (your concept)
No brief. This is the real test.

Research the market — our category, adjacent DTC brands, what's running on Meta right now. Decide what angle we should be running. Then build two static ads for it.

You choose the angle, the layout, the copy, the format. Bring us ads you'd actually put spend behind.

Good luck.

Richard
CEO Ares Underwear
----------

SKILL REQUIREMENT
VIEW OTHER JOB POSTS FROM:
SHARE THIS POST
facebook linkedin
  BENCHMARKS  
Loading Time: Base Classes  0.0013
Controller Execution Time ( Jobseekers / Job )  0.0155
Total Execution Time  0.0174
  GET DATA  
No GET data exists
  MEMORY USAGE  
1,513,952 bytes
  POST DATA  
No POST data exists
  URI STRING  
jobseekers/job/Static-Creative-Strategist-AI-Image-Generation-DTC-Menswear-Brand-1716350
  CLASS/METHOD  
jobseekers/job
  DATABASE:  onlinejobs (Jobseekers:$db)   QUERIES: 13 (0.0070 seconds)  (Hide)
0.0006   SELECT *
                                
FROM exrates
                                WHERE rate_name 
'USD-PHP' 
0.0006   SELECT *
FROM `employer_jobs`
WHERE `job_id` = 1716350
 LIMIT 1 
0.0005   SELECT *
FROM `employers`
WHERE `employer_id` = 688189
 LIMIT 1 
0.0005   SELECT COUNT(*) AS `numrows`
FROM `t_thread` `t`
LEFT JOIN `t_thread_misc` `miscON `t`.`id` = `misc`.`thread_id`
WHERE `t`.`job_id` = 1716350
AND `misc`.`idIS NULL 
0.0006   SELECT e.business_namee.logoe.websitee.rebill_datee.date_added member_datehitsDATEDIFF('2026-09-12',ej.date_added) duration_daysDATEDIFF('2026-09-12',e.rebill_date) duration_rebillej.*, e.deactivate FROM employers eemployer_jobs ej WHERE e.employer_id ej.employer_id AND
                                   ((
e.user_level >= '500' AND ej.date_added <= e.rebill_date)
                                   OR 
e.employer_id '' OR (ej.date_approved <> '2000-01-01' and DATEDIFF('2026-09-12',ej.date_added) <= 14 ))
                                   AND 
e.deactivate != AND ej.deleted AND job_id '1716350' 
0.0004   SELECT *
FROM `employer_jobs_skills` `ejs`
LEFT JOIN `skills_categories` `scON `ejs`.`skill_id` = `sc`.`id`
WHERE `job_id` = 1716350 
0.0009   UPDATE employer_jobs SET hit_counts '***Aug-25-2026=1***Aug-26-2026=414***Aug-27-2026=652***Aug-28-2026=148***Aug-29-2026=58***Aug-30-2026=44***Aug-31-2026=39***Sep-01-2026=31***Sep-02-2026=32***Sep-03-2026=24***Sep-04-2026=24***Sep-05-2026=8***Sep-12-2026=2' WHERE job_id'1716350'  
0.0007   UPDATE employer_jobs SET monthly_hits '***Aug-2026=1356***Sep-2026=121' WHERE job_id'1716350'  
0.0006   SELECT date_sent FROM jobseeker_sent_emails WHERE jobseeker_id '' AND job_id '1716350' AND status LIKE 'sent%' ORDER BY id DESC  
0.0004   SELECT *
FROM `employer_jobs_skills` `ejs`
LEFT JOIN `skills_categories` `scON `ejs`.`skill_id` = `sc`.`id`
WHERE `job_id` = 1716350 
0.0006   SELECT COUNT(*) AS `numrows`
FROM `employer_jobs`
WHERE `employer_id` = '688189'
AND `date_added` >= '2022-06-08' 
0.0004   select from teasers 
0.0003   SELECT FROM skill_categories WHERE skill_cat_id='' 
  HTTP HEADERS  (Show)
  SESSION DATA  (Show)
  CONFIG VARIABLES  (Show)