Senior ML Engineer / Research Engineer – Diffusion Models

Our client is an early-stage AI company building multimodal intelligence for interactive world creation. The goal is AI that helps anyone create playable, editable worlds with studio-grade quality in real time. This is not about offline demo clips. It is about real-time interactive content, persistence, and controllable logic.

The team is small (around ten people), talent-dense, and operates at high pace with high ownership. The work sits at the intersection of agents, diffusion, computer vision, and graphics, aimed at simulation and gaming outcomes. Seed-backed with significant runway and real upside for early joiners.

The role:
You will own and evolve diffusion model systems that can generate and edit high-quality content under real product constraints around quality, controllability, and compute. This is a builder role: research and product execution happen in the same seat.

What you will do:
Build and iterate on diffusion architectures across 2D, 3D, image, video, and audio
Design conditioning and control mechanisms across text, images, pose, layout, and other structured signals, including multimodal encoders and guidance strategies
Run large-scale diffusion training and improve quality vs. compute trade-offs using better objectives, distillation, and consistency/one-step approaches
Build controllable generation and editing pipelines (ControlNet, LoRA, IP-Adapters, inpainting, personalisation, custom subject/style tuning)
Ship real systems, not just research prototypes: build tooling, evaluation loops, and iteration cycles that improve quality and stability over time

What you will need:
2 to 10 years of relevant experience with strong signals of technical ability and speed
A track record building scalable products, not just research work that stays in a notebook
Hands-on diffusion experience including training and architecture work, not just library usage
Comfort working across disciplines, especially where computer vision or graphics meet model building
Clear builder mindset: excited by cutting-edge research and product execution in the same role

Strong plus:
Experience making models faster, cheaper, and more reliable in production
Projects touching agents and multimodal reasoning alongside diffusion
Strong academic performance or competitive programming background

For more information please reach out to [email protected]