Mahir M.

I am an AI Video, Image, and 3D Generation Specialist with 50+ completed projects and a 100% Job Success Score on Upwork. I build production ready AI generation workflows using MiniMax H3, ComfyUI, Wan 2.2/2.5, Kling, Hailuo, Flux, and other AI models. Most of my client work is for YouTube channels, UGC content, ads, product content, and other commercial projects. I handle the full workflow from prompt writing and model selection to workflow building, automation, deployment, and final output. For AI video, I work with MiniMax H3, Wan, Kling, Hailuo, and other video models. I can build text to video, image to video, reference based video, character consistency, batch generation, and automated video workflows. I also build ComfyUI workflows for image and video generation. I can connect ComfyUI to custom frontends and APIs, deploy workflows on RunPod, set up Docker containers, optimize node graphs, and handle large batch jobs. I focus on building workflows that are clean, reusable, and easy to run again. The goal is to give you a workflow you can keep using without needing me to rebuild it every time. I also work with LoRA training. If you need a specific character, product, person, or style trained into a model, I can prepare the data, train the LoRA, test it, and add it to your workflow. For 3D generation, I work with Hunyuan 3D, Trellis 2, Meshy, Tripo, and Hyper3D Rodin. I can turn a reference image, render, or text description into a usable 3D model with PBR textures for Blender, Unreal Engine, product visualization, virtual fashion, and game assets.