__tl;dr: fullstack engineers, $135k-270k, relocation to London __

A year has passed since the post where we wrote about Apollo research — back then, the fresh Claude 3.7 understood it was being tested in 23% of cases. Now that sounds almost touching. Here's what has happened in this past year:

🟢 Apollo has officially partnered with OpenAI! And they are conducting joint research on scheming — that is, on how models deceive, trick, and hide their intentions; 🟡 Apollo CEO Marius Hobban has been included in the top 100 most influential people in AI by Time magazine for 2025; 🔴 The latest models guess they are being tested in most cases — OpenAI and Anthropic are now directly writing about this, admitting they have almost no way to measure how honestly they behave 🤷‍♂️

In general, coming up with new ways to evaluate models is becoming only more important.

To continue doing this, Apollo needs **Full Stack Software Engineers.

**They will be responsible for creating a wide variety of tools for research — something like an IDE for evaluations!

**For example: **⭐️ LLM search that finds interesting fragments in evaluation transcripts; ⭐️ streaming results of ongoing experiments; ⭐️ multi-user log editing that will automatically update related metrics.

Requirements: 5+ years of strong full-stack experience (frontend is as important as, or slightly more important than, backend).

💪 Strong means, for example: 🔵 you have led a product or a major component for over a year; 🔵 or you have created a popular open-source tool, built the entire stack in a startup; 🟢 or you have rapidly grown in a large company to a role with significant responsibility.

Tech stack — Python and React.

They offer relocation to London, pay **$135k-270k, **provide free lunches and dinners at the office, unlimited PTO, and a development budget.

Recommend your best full-stacks to Diana @yourdivna — and we remind you that we pay a referral bonus if we hire someone based on your recommendation 🤍