Prompt engineer for Gemini / Nano Banana (Character generation + Face replacement)

Budget: 100 USD Deadline: 7 days

Banana! 😀 Hello! I have practical experience in prompt engineering for multimodal models like Google Gemini/Imagen, setting up pipelines in n8n/Python, and creating photorealistic generations without AI artifacts.

🛠 How I will do it:
1. I will develop a flexible prompt template with dynamic variables for creating character types while preserving nationality, features, and angles (70-80% similarity).
2. I will create a prompt for Inpainting/Face Swap, precisely adjusting the new face to the lighting, shadows, and geometry of the original scene without a masking effect.
3. I will test API calls, select optimal hyperparameters (temperature, top_p, guidance), and prepare ready JSON templates for n8n.

🎥 Portfolio: Freelancehunt

I am ready to complete a test task and discuss the details!

Budget: 100 USD Deadline: 2 days

Banana, Timur. I immediately see the task: you need not just prompts, but stable templates for the API that maintain similarity, realism, and do not break the scene. I have worked with content automation and understand how important repeatability, parameter control, and a clean result without artifacts are. I can create 2 universal prompts for Gemini/Nano Banana, provide a variable structure for JSON/n8n, and a test run scheme for 2–3 people and face swap. I am ready to discuss the details and demonstrate the approach on your case.

Budget: 100 USD Deadline: 3 days

Hello.
I will complete the work quickly and efficiently.
Extensive practical experience.
Google Drive
Google Drive

Budget: 100 USD Deadline: 14 days

Good day.

I understand the task as the development of two stable parameterized templates, rather than two one-time text requests, which can be used with different photographs through the Gemini API or n8n without manual rewriting.

To generate variable types, I will build a prompt with controlled variables:

— level of similarity and variability;
— gender, age, and general anthropometric characteristics;
— preservation of race, national features, eye color, and face shape;
— modification of appearance details without fully copying the person;
— natural skin texture without a plastic effect;
— realistic lighting, optics, depth of field, and detail.

For face replacement, I will separately outline the requirements for pose, facial expression, gaze direction, lighting, skin tone, perspective, and background preservation. I will add negative constraints against distortions, double facial features, unnatural eyes, waxy skin, and artifacts.

Before delivery, I will conduct testing on photographs of varying quality, gender, age, lighting, and angles. I will also prepare a structure of variables for the API/n8n and a brief instruction on setting the level of similarity.

My approach is favorably different from most performers in that I will work not by randomly selecting words, but through a test matrix, controlled parameters, and repeatability of results. I will execute quickly, efficiently, and within a moderate budget.

I suggest discussing examples of desired results, API format, and acceptance criteria in private messages. Expand

Budget: 150 USD Deadline: 2 days

Banana. I will develop two universal JSON templates for n8n (Imagen 3 / Nano Banana API) that ensure result stability without manual adjustments.

What you will receive:
1. Phase 1 Template: Generation of 10 unique characters with 75% similarity. The prompt captures ethnicity, skull geometry, and eye color, changing only micro-asymmetry and skin details.
2. Phase 2 Template: Universal Inpainting for seamless face replacement. The algorithm automatically adapts the new face to the lighting, shadows, and grain of the original scene, completely blocking changes to the background or hairstyle.

Technical details:
— Prompts optimized for "Crystal Realism" (Sub-surface scattering, micro-pores, zero AI blur).
— Templates fully ready for insertion into the HTTP Request node.
Ready for testing right now: send a photo, and I will demonstrate the generation of characters and their seamless integration into your scenario.

Budget: 100 USD Deadline: 2 days

Banana

Hello!

Your project is exactly the kind of prompt engineering challenge I enjoy. I have experience designing structured, reusable prompts for multimodal AI workflows and understand that the goal here is not simply to achieve a good result once, but to create stable, template-driven prompts that work reliably in automated pipelines.

I can help develop:

* A universal template prompt for generating multiple realistic face variations from a reference image with configurable similarity.
* A universal template prompt for image editing/inpainting workflows that integrates cleanly into an automated pipeline.
* Dynamic prompt templates with variables suitable for JSON, n8n, Make, or Python-based workflows.
* Clear documentation describing the purpose of each variable, recommended parameter ranges, and integration guidance for automated execution.

My approach focuses on repeatability and consistency rather than manually tuning prompts for individual images. During development, I test multiple scenarios using different input images to improve robustness before delivering the final templates.

I am comfortable participating in a test process and iterating based on the results to achieve the best possible quality within the capabilities of the selected model.

Please note that the achievable behavior and supported parameters depend on the specific Google model and API version being used. I will tailor the templates and recommendations to the exact API you plan to integrate.

I am available to start immediately and would be happy to discuss your workflow, automation pipeline, and success criteria before we begin.

Thank you, and I look forward to working with you. Expand

Budget: 100 USD Deadline: 1 day

"Banana" Ready to develop two universal, template-based prompts

Budget: 100 USD Deadline: 1 day

Banana 🍌

Hello! I am interested in your project.

I work with Gemini, Nano Banana, Imagen, and other AI models, regularly writing prompts for realistic generation and face replacement. I understand how to achieve the most natural results and prepare templates for automation through API/n8n.

I am ready to complete a test task right now. If you are satisfied with the result, I would be happy to start our collaboration.

Budget: 100 USD Deadline: 1 day

Banana! I welcome an excellent and clear task for building a fault-tolerant AI pipeline. I work with Gemini / Imagen 3 Nano Banana via API and know how to achieve precise facial geometry and realistic microtexture of pores, imperfections, natural light and shadow patterns without the mask and blur effect. I am ready right now to do a test task to generate 2–3 variant faces based on your source and perform a test face replacement in the scene.

Send the source photo and target frame in the chat — I will show the result!

Budget: 100 USD Deadline: 3 days

Hello!

I am interested in your project. I am ready to promptly take on the task and discuss all the details.

I work with modern AI tools for generating images, videos, and creative content.

Please write to me in private messages.

Budget: 100 USD Deadline: 3 days

Banana. Prompt stage 1 should capture race, skull shape, and eye color through descriptors, not pixel binding to the reference, then 70-80% similarity is regulated by one parameter similarity_weight.

Stage 2 is inpainting based on the face mask: negative on the background, clothing, and pose plus adjustment for light and color of the frame. I will create both as JSON templates for n8n with parameters temperature and top_p.

I am ready to confirm in practice: 2-3 types and a test face replacement. Access through AI Studio or Vertex AI?