Diffusion Specialist

AutoHDR•Austin, TX

2h•Onsite

About The Position

We are looking for unusually strong engineers who can ship broadly across the stack and also bring deep expertise in diffusion models, generative modeling, and image synthesis systems. We want builders who can contribute to product features, backend systems, internal tools, debugging, and AI workflows, while owning and advancing the core diffusion-based systems that power our image editing capabilities.

Requirements

You are a high-agency builder who likes ownership
Experience as a founder, CTO, founding engineer, or having owned a core ML or generative system
Deep experience with diffusion models, generative modeling, or advanced computer vision systems
Work is connected to real product systems and users
Experience with zero-to-one model or system building
Ownership of a core generative model or image system in production
Strong understanding of diffusion models and generative techniques
Experience improving model quality (consistency, realism, control, artifact reduction)
Experience with training, fine-tuning, or adapting large models
Ability to ship models into real-world systems with real users
Experience with real shipped products with real users
Exceptional open-source, research, or competitive ability when it clearly reflects real technical horsepower

Responsibilities

Design, train, and improve diffusion-based image editing systems
Build and optimize training pipelines for generative models
Develop high-performance inference pipelines for image generation and editing
Improve model quality across realism, consistency, controllability, and style fidelity
Optimize diffusion systems for speed, cost, and scalability in production
Collaborate with product and engineering to ship model improvements to users
Implement / Improve Systems For: User-facing AI image editing tools powered by generative models
Implement / Improve Systems For: Property manipulation tools (virtual staging, re-lighting, object edits, etc.)
Implement / Improve Systems For: Style consistency and controllable editing systems
Implement / Improve Systems For: High-throughput image generation and editing pipelines