Diffusion Models Researcher

About Us

Martian Lawyers Club (MLC) is pushing gameplay past its current limitations to unlock new levels of video game personalisation. We want our games to do what is only possible in the tabletop RPG world: indefinite personalisation and endless creativity. We believe that players shouldn’t be forced to pick between the immersive power of video games and the ability to make unconstrained choices.

Our proprietary auto-modding platform uses recent breakthroughs in machine learning to allow games to modify themselves according to player desires — tapping directly into the DNA of a game and modifying that as part of the core gameplay loop. We are now taking this further with a local, real-time video generation models that visualise emergent gameplay states, and we need you to help build it.

Responsibilities

What you’ll do from day to day:

• Research, prototype, and iterate on diffusion model architectures suitable for real-time, controllable video generation on consumer hardware.
• Design and implement conditioning mechanisms that allow game state and player actions to guide video synthesis.
• Develop and maintain efficient inference pipelines, including quantisation, distillation, and hardware-aware optimisations.
• Collaborate closely with the game team to integrate the generation model into our live game loop.
• Publish findings internally and contribute to the broader research community where appropriate.
• Stay current with the rapidly evolving landscape of video generation, world models, and related ML research.

About You

Desired qualifications and experience:

• Strong background in deep learning research, with a focus on generative models (diffusion models, flow matching, or similar).
• Hands-on experience training and fine-tuning video or image generation models.
• Solid understanding of efficient inference techniques: quantisation, distillation, model pruning, and hardware-specific optimisation (CUDA, TensorRT, etc.).
• Proficiency in Python and PyTorch (or JAX).
• Ability to read, implement, and critically evaluate recent ML research papers.
• Strong communicator who can bridge the gap between research and product.
• A genuine interest in video games is a big plus.

It’s okay if you don’t tick all the boxes. We particularly encourage applications from candidates who are underrepresented in the gaming industry.

What We Offer

We’re big believers in personal responsibility. You will be given the freedom to make your own choices, as you often will be the person who understands your area best. But with great responsibility also comes great power! Joining us this early means you are sure to have a massive impact on the direction the company takes.

Our work environment is dynamic, a mixture of in-person and remote, and gives you the space to learn from your mistakes (of which we expect many!). Your day-to-day work will continuously evolve as the field does, so expect to be challenged and to grow rapidly.

Application Process

Please send your CV to hiring@martianlawyers.club with a subject line of “MLC Diffusion Models Researcher”. Include links to any relevant research, papers, or projects in the body of the email. We look at every application we receive and aim to reply as quickly as is feasible.

The interview process consists of:

• A 15-minute introductory call with the founder.
• A take home task that takes about 3 hours to complete.
• A 1-hour online interview.
• A final, onsite interview in one of our offices.

After this we send our final offer decision.