Helium is building self-improving software.

We previously led AI & Growth teams at Uber, Meta, and Amazon, and we saw how long it takes to go from idea to experiment for these large companies.

We’re building Helium to automate the end-to-end process of generating and experimenting on new in-app experiences.

We’re starting by automating generating onboarding flows for consumer apps, then will expand into other experiences across the app. We work with some of the largest, fastest-growing consumer apps, serving Helium experiences to millions of users worldwide.

Our momentum:

• Launched publicly in June '25. Growing X00% month over month since.

• Working with leading consumer companies, serving experiences to millions of users daily.

• Backed by >$5M from Y Combinator, Google’s AI fund, and leaders from Uber, OpenAI, Meta, and more.

Benefits:

• Build on the newest AI models, even before they're released publicly
• Work with a stellar hard-working team at a prime office in SF
• All meals (Breakfast, lunch, dinner) covered at the office
• Create the future of software

Helium is building self-improving software — AI that continuously creates and tests new software experiences.

Backed by Y Combinator, Gradient, and leaders from Uber, OpenAI, Meta, and more. We’ve raised $5MM+ to build our vision.

What You’ll Do

• You will be building an AI paywall editor that serves millions of customers everyday, including touching the front-end Next.js application and back-end high performance Python server.
• Integrating the latest coding models, APIs, and techniques into the editor to improve generation performance.
• Writing and deploying Python code that serves hundreds and thousands of requests each second globally.
• Work cross functionally with the founders, and directly with customers, to design, implement, and deliver the best experiences!
• Wearing a lot more hats than listed above 👷.

What We’re Looking For

• Keen problem solving skills: you are able to cut down ambiguous problems down to first principles and deliver a working solution.
• Strong programming/full-stack foundation: You’re comfortable working on both backend code as well as frontend React code.
• Experience working directly on user-facing products: you are able to figure out and understand what the user needs, including talking to customers
• Stellar cross-functional communication skills: you are able to adapt to our ways of working and be an efficient/effective communicator.
• Owning features end to end: you’ll write your own PRDs/quick specs when necessary, and you test your features end to end to make sure they work and satisfy the requirements

Nice To Haves

• Experience with ML/AI/LLM integration in applications
• Background working with design tools, chat experiences, or vibe coding tools
• Previous experience founding or working at a high-growth startup
• Mobile engineering/mobile experience is a plus

Reasons This Role Might Not Be A Fit

• You’re uncomfortable with ambiguity, or need super structured requirements
• You prefer to only one part of the stack, and aren’t willing to work across everything
• You’ve never worked on systems with many moving parts: Our software has many different moving parts that you will be required to ramp up on

Requirements

• You have 1-4+ years of engineering experience, with a strong track record of shipping real products.

  🔴 Note 🔴

  • This is a full-time, in-person role based in San Francisco - we work from the office 6 days a week (Mon-Sat)
  • You must be based in the Bay Area or willing to relocate before starting.
  • We require US work authorization, but are open to H1B // O-1 visa sponsorship for truly exceptional candidates.

Our stack:

• Next.js (React/Typescript), Zustand
• Tailwind/Shadcn
• Supabase
• Python/Fly.io
• Mobile: Swift, Flutter, Expo

Here are the technical problems you're solving:

• RL for software personalization - dynamic assignment of price, paywalls, and experiences to different users based on what improves metrics

• Automated a/b testing - building systems that automatically chooses and runs new paywall experiments to run based on analytics + codegen models

• AI editor - think Bolt/Lovable for paywalls

• Billion scale content serving - we're building a python service to process thousands of QPS that serving assets to millions of phones daily → soon to be billions

• Mobile webview optimization - building a webview layer that feels native across 4 mobile stacks