AI Content Virtual Assistant — Higgsfield AI Avatar & Product Image Generation

We are looking for a reliable, detail-oriented Virtual Assistant who is highly skilled at creating **hyper-realistic AI images using Higgsfield**.

The primary responsibility of this role will be creating and maintaining a **consistent, realistic female AI avatar** that can be used across a large number of images, poses, environments, products, and before/after scenarios.

This is intended to be a **long-term position** with significant potential for the workload and compensation to increase based on performance.

Compensation

Starting Pay: $50 USD per week

There is significant potential for compensation to increase over time based on the quality, speed, consistency, and volume of your work.

We are looking for someone interested in learning our exact process and growing with the operation rather than completing a short-term project.

REQUIRED TOOL — HIGGSFIELD

You will be required to use Higgsfield for this position.

Experience with Higgsfield is strongly preferred. If you have experience creating consistent AI characters/avatars and realistic product images using Higgsfield, please make this clear in your application.

You should be comfortable working with multiple reference images, writing and adjusting prompts, generating variations, and troubleshooting generations until they meet our quality standards.

What You'll Do

Your primary responsibility will be creating hyper-realistic AI-generated photos of a consistent female avatar.

I will provide you with:

* A picture/account showing the general type of girl and aesthetic I want you to use as inspiration
* Examples of AI-generated images I have successfully used in the past
* Reference images showing poses, settings, products, and concepts
* Actual product reference images
* Examples of how I want the avatar to hold/display products
* Instructions for the specific images I need

You will then use **Higgsfield** to create an original AI avatar and generate realistic images of that same girl.

The AI avatar should **not simply copy the identity of the reference person**. The reference is intended to communicate the general appearance, aesthetic, content style, poses, and level of realism I am looking for. The final avatar should be a distinct AI-generated person.

Avatar Consistency

Consistency is extremely important.

Once we establish the AI avatar, she should remain recognizable as the same person across different generations

Her core identity should remain consistent, including:

* Face and facial structure
* Hair
* Skin tone
* General body type
* Age appearance
* Overall aesthetic

I should be able to look at multiple generated photos and immediately recognize that they are all supposed to be the same girl.

At the same time, the images should not look repetitive or artificially identical. She should be able to appear naturally in different poses, outfits, rooms, lighting conditions, camera angles, and situations while maintaining the same identity.

Before & After / Pain-Point Images

A major part of this position will involve creating **before-and-after images** involving different cosmetic or appearance-related pain points.

For example, I may ask you to create images where the avatar has:

* Yellow/stained teeth
* White teeth
* Under-eye bags/dark circles
* Clearer/brighter under-eyes
* Acne or blemishes
* Clear skin
* Razor bumps
* Other visible cosmetic concerns

You may be asked to create a **BEFORE image** showing the pain point and an **AFTER image** showing an improved appearance.

The most important requirement is that the person still looks like **the exact same AI avatar in both images**.

The pain point should look realistic and naturally integrated into the photo. It should not look exaggerated, fake, painted on, or obviously AI-generated.

For example, if creating a yellow-teeth before image and a white-teeth after image, the girl's identity, facial structure, hair, environment, and other important details should remain consistent. The primary visual change should be the requested pain point.

Product Images

Another major part of this position will be creating realistic product images featuring the AI avatar.

I will provide you with a clear reference image of the actual product, as well as examples showing the type of image I want created.

For example, I may ask you to generate an image where the AI avatar is:

* Holding the product toward the camera
* Holding the product naturally in her hand
* Showing the product while taking a selfie
* Using the product
* Posing with the product
* Displaying the product in a realistic UGC-style photo

The product must remain **extremely accurate to the reference image.**

This includes:

* Product shape
* Size and proportions
* Packaging
* Colors
* Logos
* Branding
* Labels
* Product design
* **All visible text**

### PRODUCT TEXT MUST BE LEGIBLE

This is extremely important.

AI-generated products often contain distorted, misspelled, random, or unreadable text. **This will not be accepted.**

If the real product says something on its packaging, the generated product should accurately display that same text wherever it is clearly visible.

Product text should:

* Be spelled correctly
* Be readable
* Match the actual product
* Have realistic typography
* Not contain random AI-generated letters or words
* Not appear warped or distorted
* Not immediately reveal that the image was generated with AI

If Higgsfield generates an otherwise good image but the product or text is inaccurate, you will be expected to **fix the issue or regenerate the image before submitting it.**

The final result should look like the girl was **actually photographed holding the real product.**

Reference Images

You will frequently receive reference images that demonstrate what I want.

These references may show:

* A pose I want recreated
* A specific camera angle
* A selfie or mirror-selfie style
* A facial expression
* A room/background
* A product
* How the product should be held
* A before/after concept
* A specific pain point
* The general aesthetic of the content

Your job is to understand **what elements of the reference matter** and recreate the concept using our established AI avatar.

You should NOT simply copy the original creator/person. The goal is to recreate the useful content concept while using our own consistent AI-generated avatar.

## Image Style

The final images should be **extremely realistic**.

In many cases, the desired result should look like a normal photo taken on an iPhone rather than a professional photoshoot or obvious AI generation.

We want:

* Natural skin texture
* Realistic hair
* Realistic teeth
* Realistic hands and fingers
* Natural body proportions
* Realistic lighting and shadows
* Believable backgrounds
* Natural poses
* Realistic facial expressions
* **Highly accurate products**
* **Perfectly legible product text**
* Realistic imperfections
* Natural iPhone/selfie image quality

We DO NOT want images that look overly smooth, overly perfect, heavily edited, plastic, CGI-like, or obviously AI-generated.

## What We're Looking For

* Experience using **Higgsfield**
* Experience creating hyper-realistic AI people
* Ability to maintain the same AI character across many generations
* Experience generating realistic AI product images
* Ability to maintain accurate branding and product details
* Strong attention to text, logos, labels, and packaging
* Experience working with reference images
* Strong prompting skills
* Ability to troubleshoot poor AI generations
* Excellent attention to detail
* Ability to create realistic before/after variations
* Ability to identify and correct obvious AI mistakes
* Ability to follow detailed instructions
* Ability to work independently
* Ability to consistently meet deadlines
* Reliable internet connection
* Good written English and ability to understand spoken English

Experience creating realistic AI UGC/social-media content is a major plus.

# REQUIRED 6-IMAGE APPLICATION TEST

**To be considered for this position, you must complete a 6-image trial using Higgsfield.**

The purpose of this test is to determine whether you can create a hyper-realistic AI avatar, maintain the same person's identity across multiple images, realistically modify specific features, and accurately incorporate real products.

## AVATAR STYLE REFERENCE

REFERENCE ACCOUNT / PERSON:

[ ---------- ]

Use this reference to understand the **general aesthetic and type of avatar** we are looking for.

Do NOT simply duplicate the exact person's identity. Create an original AI-generated girl inspired by the general aesthetic.


**PREVIOUS EXAMPLE:**

[will provide over text]


## Trial Instructions

First, use **Higgsfield** to create one hyper-realistic female AI avatar.

You will then create **6 total images of the SAME AI GIRL.**

### IMAGE 1 — Normal Avatar

Create a realistic photo of the avatar with a natural appearance.

### IMAGE 2 — Yellow Teeth / BEFORE

Create a realistic image of the avatar showing noticeably yellow/stained teeth.

### IMAGE 3 — White Teeth / AFTER

Create the corresponding AFTER image showing the **same girl** with naturally white teeth.

Images 2 and 3 should clearly look like before-and-after photos of the same person.

### IMAGE 4 — Under-Eye Bags / BEFORE

Create a realistic image of the avatar with noticeable but believable under-eye bags/dark circles.

### IMAGE 5 — Improved Under-Eyes / AFTER

Create the corresponding AFTER image of the **same girl** with the under-eye area visibly improved.

Images 4 and 5 should clearly look like before-and-after photos of the same person.

### IMAGE 6 — PRODUCT IMAGE

Create a hyper-realistic image of the **same AI girl holding the provided product**.

Use the product reference and example provided above.

This image will be graded heavily on **product accuracy**.

The product should look like the real product was physically placed into the girl's hand and photographed.

**All visible branding, logos, labels, and text must be accurate and legible. There should be no obvious AI-generated text, misspellings, distorted logos, warped packaging, or other AI artifacts.**

Her hand and fingers must also interact naturally with the product.

## What We Will Grade

Your trial will primarily be judged on:

**Avatar Consistency → Realism → Product Accuracy → Text/Logo Accuracy → Before/After Accuracy → Attention to Detail → Ability to Follow Instructions → Overall Image Quality**

The biggest questions we will ask are:

**"Do all six images convincingly look like real photos of the same girl?"**

and

**"Does the product look real enough that someone would not immediately know it was AI-generated?"**

## How to Apply

Please submit:

* A brief introduction
* Your experience with AI image generation
* Your experience specifically with **Higgsfield**
* Your experience creating realistic AI product images
* Examples of realistic AI people/avatars you have created
* Examples of AI product images you have created, if available
* Your daily availability
* **All 6 completed Higgsfield trial images**

Please include the word **"AVATAR"** at the beginning of your application so we know you read the full job description.

**All 6 trial images must be submitted with your application. Applications without the completed trial will not be considered.**

Candidates who demonstrate the strongest combination of **realism, avatar consistency, product accuracy, readable text, attention to detail, and Higgsfield skill** will be contacted about the ongoing position.

**Starting compensation is $50 USD per week, with the potential to increase significantly based on performance, reliability, quality, and increased workload.**