Open roles

Content Engineering Intern

Team
Engineering
Type
Internship, 4 weeks
Location
London (Hybrid, 4 days in-office)

About 36 Labs

36 Labs is a seed-stage startup founded out of UCL’s machine learning department. We bootstrapped to 10 people, and now we're on the next leg of growth.

AI can generate almost any form of media, but it still can't judge what captures human attention. AI-generated content on social platforms is growing fast, every creative workflow will be agent-assisted or fully agentic. Our work is improving how models reason about creative decisions, learning from how audiences respond to content.

The role

You'll push the frontier of what our creative harness can make. You'll find out which combinations of image, video, voice and music models, prompts and editing techniques produce the best content, and turn them into methods the harness runs in production. Videos made by our system have passed 300M+ views on TikTok, and the bar keeps rising.

This is for someone young and hungry. Fresh grads are welcome, and you don't need years of experience. You need taste, speed, and a habit of making things. Ideally you've already grown a social media account or run content campaigns yourself, so you know what it takes to make people stop scrolling. New models and workflows come out every week, and you'll be the first here to try them. When a creator posts a workflow that's going viral, you'll reverse-engineer it, improve it and make it ours.

It starts as a 4-week paid internship. You'll ship real work into production from week one, and if it's a fit, we'll convert it into a full-time role.

You'll work closely with:

  • the CTO, on direction
  • harness research engineers, on how the harness plans and generates
  • backend engineers, on the platform your methods run on
  • our agency team at Outer Circle, who run live campaigns for labels and brands and put your work in front of real audiences

What you'll do

  • Test new image, video, voice and music models as they release. Find the chains of models and steps that beat what we use today, and know when a cheaper or faster tier is good enough.
  • Run experiments end to end. Generate enough samples to judge fairly, compare outputs side by side, and record what worked.
  • Reverse-engineer viral formats and creator workflows from X, TikTok and Instagram into repeatable methods within days.
  • Turn the winners into methods the harness runs: which model, which provider, what it costs, and how it fails.
  • Build small tools that keep us ahead, like trackers for new models, workflows and formats that are taking off.
  • Work with Outer Circle on creative concepts and samples for new clients and campaigns, and use their feedback and live results to decide what to scale and what to cut.
  • Hold the quality bar. Every format we ship should be something that could get millions of views.

What we're looking for

  • You've run a social media account, or campaigns on platforms like Whop, and grown an audience or made content that performed.
  • You're scrappy and high agency. You find a way to get things done without waiting to be told how.
  • You move fast. You can go from "someone posted this workflow" to a working version the same day.
  • A portfolio of image or video work made with AI that you produced yourself and that doesn't look like slop.
  • Hands-on experience with generative media tools and models (ComfyUI, fal, Higgsfield, Kling, Veo, Seedance, Flux and the like). You've chained models together, not just prompted one.
  • You can code in Python or TypeScript well enough to automate your own workflows.
  • Taste and an understanding of short-form: hooks, pacing, captions, editing, and why one video gets watched while another gets scrolled past.
  • You're extremely AI-pilled: you work with models every day and know the latest releases, techniques and recipes as they land.
  • Projects or content you've made and can walk through in depth: what you tried, what failed and why the final version works.

Nice to have

  • A post that went viral or an account with a real following.
  • Fine-tuning or LoRA training for image or video models.
  • A background in video editing, motion design, film, photography or design.
  • Building evals or comparison tools for generative media.

Our stack

Python, TypeScript, Postgres, LLM and generative media APIs, open-weight models.

We encourage you to apply even if your experience doesn't match every line above.