Generative Image AI Engineer
hace 8 horas
Barcelona
Rendair AI is an AI rendering platform built for architects and designers, providing tools to create high-quality images, edit renders, upscale outputs, and generate videos through an intuitive and user-friendly interface. The platform enables professionals to transform ideas into final visuals within minutes, without complex setups or external tools. Trusted by thousands of architects, design studios, and creators worldwide, Rendair AI delivers fast and reliable results every day. We are focused on accelerating creative workflows, improving visual quality, and making advanced AI image generation accessible to professional users. We are looking for a hands‑on Generative Image AI Engineer with deep practical experience in image generation, diffusion models, and image‑to‑image workflows. This role is specifically for someone who has worked directly with generative image models: testing and comparing models, fine‑tuning or adapting image workflows, improving image quality and consistency, connecting model APIs or endpoints, and turning image AI capabilities into real product features. We are not looking for a general ML profile or a research‑only candidate. The ideal person is highly hands‑on with image models and visual AI workflows, including tools such as Stable Diffusion, SDXL, Flux, ControlNet, LoRA, IP‑Adapter, ComfyUI, inpainting, masking, upscaling, and image‑to‑image generation. You should be technical, curious, fast‑moving, and comfortable working directly with generative image models, Python, PyTorch, APIs, endpoints, and AI tools to improve product quality and user experience. You do not need to be an architect or have experience with architecture software. What matters most is your ability to work hands‑on with AI image models and practical generative workflows. Build, test, and improve generative image workflows using open‑source and commercial AI tools. Work with models and tools such as Stable Diffusion, SDXL, Flux, ControlNet, LoRA, IP‑Adapter, ComfyUI, RunPod, Replicate, Hugging Face, OpenAI, Claude, Gemini/Nano Banana, and similar platforms. Improve workflows for text‑to‑image, image‑to‑image, inpainting, masking, upscaling, enhancement, prompt‑based editing, style control, reference control, and output consistency. Experiment with different conditioning methods such as depth, canny, lineart, segmentation, reference images, masks, and style adapters. Evaluate new open‑source and commercial image models to understand whether they can improve Rendair’s product. Train, fine‑tune, or customize image models where useful, including LoRAs and other lightweight adaptation methods. Prepare and curate datasets, test training configurations, compare checkpoints, and document quality differences. Measure and compare model outputs based on image quality, prompt adherence, consistency, controllability, speed, cost, and reliability. Understand when fine‑tuning is necessary and when better prompting, conditioning, model routing, or workflow design is a better solution. Connect AI models, tools, and third‑party providers through APIs and endpoints. Build prototypes that combine multiple AI tools into usable workflows. Work with engineers to productionize successful AI workflows. Read API documentation, test endpoints, debug integrations, and clearly document how workflows should operate. Help define the input/output structure for model workflows, including images, masks, prompts, parameters, job status, and final outputs. Use LLMs and AI agents to improve prompt generation, user guidance, image review, quality control, and workflow automation. Experiment with OpenAI, Claude, Codex, Cursor, and other AI coding or agent tools to accelerate development. Help design AI‑assisted experiences that make the product easier, faster, and more powerful for users. Improve image output quality and generation consistency. Help reduce generation time and cost where possible. Compare different model providers, workflows, inference options, and deployment approaches. Balance quality, latency, controllability, and cost based on product needs. • Strong Python experience., • Strong working knowledge of PyTorch., • Hands‑on experience with diffusion‑based image generation models such as Stable Diffusion, SDXL, Flux, ControlNet, LoRA, IP‑Adapter, ComfyUI, or similar tools., • Experience building image generation workflows beyond simple prompting, including image‑to‑image, inpainting, masking, upscaling, conditioning, reference control, and prompt‑based editing., • Experience training, fine‑tuning, or adapting image models for practical use cases, especially LoRAs or similar lightweight model adaptations., • Experience preparing datasets, testing training configurations, comparing checkpoints, and evaluating output quality., • Experience connecting AI models, APIs, endpoints, model providers, or automation workflows., • Experience using platforms such as RunPod, Replicate, Hugging Face, OpenAI, Claude, Gemini, or similar AI tools., • Ability to evaluate models and workflows based on image quality, consistency, controllability, latency, cost, and reliability., • Comfortable reading technical documentation, testing new models quickly, and documenting workflows clearly., • Strong practical problem‑solving ability and willingness to experiment., • Product mindset: you care about turning AI capabilities into useful, reliable user experiences., • Experience with ComfyUI workflow design and deployment., • Experience with Flux, SDXL, Qwen Image, WAN, video generation models, or other recent generative image/video models., • Experience with image processing, segmentation, depth maps, normal maps, lineart, canny, or 3D‑aware workflows., • Experience with inference endpoints, RunPod, Replicate, Hugging Face endpoints, Modal, GCP, AWS, or other GPU‑based infrastructure., • Experience with backend APIs, queues, webhooks, async jobs, or model versioning., • Experience with AI coding tools such as Claude Code, Cursor, Codex, or GitHub Copilot., • Experience building internal AI tools for creative, design, product, or content teams., • Background in visual design, 3D rendering, gaming, VFX, architecture, interior design, or real estate. Visualization is a plus, but not required. #J-18808-Ljbffr