Technical Artist - Language & Prompt Systems

We’re looking for a Technical Artist – Language & Prompt Systems to join our team at Luma. This isn’t a traditional “artist” role. You’ll be deep in the world of LLMs, prompt language, and multimodal AI — exploring how subtle shifts in phrasing shape visual output. You’ll be crafting, testing, and iterating on prompts that power some of the most advanced generative models on the planet.

If you love the intersection of language, creativity, and precision, this is your playground.

What a Typical Week Might Look Like

• Tuning and testing dozens of prompt variations to find the most expressive results

  Debugging strange or surprising model outputs and proposing fixes

  Writing internal guides to help teammates understand and reuse your best prompts

  Curating output examples for training datasets or product showcases

  Collaborating with researchers to design language evaluations or edge-case tests

  Building light tools or scripts (with help) to streamline prompt QA workflows

  Annotating and captioning visual samples with absurd attention to detail

  Spotting a new prompting trick on Twitter, testing it, and making it better

Your Background Looks Something Like

• You’ve spent serious time with tools like GPT-4, Claude, Midjourney, or Luma

  You understand how small language tweaks lead to big output changes

  You’re detail-obsessed and systematic — tracking, testing, and documenting your work

  You’re comfortable using AI tools to move fast, prototype, and iterate

  You have and can confidently write Python scripts, with or without the help of AI coding

  You care about quality, clarity, and the craft of communication

  You enjoy doing the same thing five different ways — to see what works best

You Might Really Thrive Here If…

• You’ve ever rewritten a sentence 20 times just to make it feel right

  You get joy from labeling, organizing, and curating messy creative output

  You love GenAI and already spend your free time exploring prompts for fun

  You’re as curious about language quirks as you are about model behavior

  You think of prompt engineering as part science, part poetry

  You’re energized by small-scale chaos and rapid iteration

Some of the Cool Things We’re Working On

• Prompt systems that drive high-fidelity video generation from text

  Language interfaces that shape model behavior, tone, and visual output

  Lightweight QA pipelines to catch weirdness before it ships

  Curated libraries of prompt recipes, edge cases, and style guides

  Interfaces for creators to shape model outputs more intuitively

  Prompt workflows that scale across research and product teams

  This role is remote-friendly, but we love collaborating in person when possible. Bay Area preferred. You’ll be joining a small, fast-moving team where your ideas and experiments directly shape the product — and the future of generative AI itself.

  About Luma

  At Luma, we’re building multimodal AI to expand human imagination and capabilities. Our mission is to create systems that can see and understand, show and explain, and eventually interact with the world to effect change. Our flagship product, Dream Machine, generates high-fidelity video from simple text — and we’re just getting started.

  We’re a product-focused, high-velocity team of researchers, engineers, and creatives building foundation models and beautiful user experiences. If you love building things that feel magical — and you want to shape the future of AI — we’d love to meet you.

  Originally posted on Himalayas