Software Engineering Intern

Description

The company is building a shared context layer for recursive companies, meaning teams whose agents learn from the company's own history.

The product has three parts: a data lake that collects what a team produces with AI, an ontology that turns that data into typed, trustworthy knowledge about the company's current state, and a loop that helps people and agents improve from it.

You would work directly with the founders on a live product. Nothing here is a sandbox: you own a real piece, ship it to real users, and watch how they use it.

Responsibilities

• Receive a scoped project in week one and have your work live in production by the end of the internship.
• Own one or two areas, scoped to what you can ship during the internship.
• Ingestion: build a new source connector that pulls from AI chat apps and coding agents, as well as meeting recorders, notes apps and email. Capturing data across providers is the hardest unsolved problem in this space.
• Search and retrieval: make thousands of sessions and documents searchable in milliseconds, both locally and in the cloud.
• Self-improving agents: measure skill use, mine agent traces for failures and repeated workflows, and propose edits to a team's skills that people can review.
• Agent-facing surfaces: work on the CLI, MCP connector and plugins so any AI tool can reach a team's accumulated context.
• Ontology: add derived fields, resolve conflicts, and keep a company's live state current as new work comes in.
• Product polish: improve the experience from first launch through the features people use every day.

Requirements

• You have shipped real software, such as a side project, a hackathon build people actually used, or a tool you still use daily.
• You have taken something from nothing to working software at least once.
• You can pick up a new stack quickly. The team ships native desktop apps on macOS, Windows and Linux, a CLI and background daemon, agent integrations and a cloud backend, written in several languages. You do not need experience in all of them.
• You use AI tools by default, such as Claude Code, Cursor or Codex, and work faster because of them. Under pressure, you hand code to an agent rather than stare at it.
• You have the instinct to ship before something is perfect.
• You work with autonomy. You get context, feedback and quick answers from the founders, but no detailed spec: you get a problem and work out the solution.
• You have opinions about the products you use and can say what is wrong with them and how you would fix it.
• You are comfortable with ambiguity, since priorities can change between week 1 and week 6.
• You do not need a specific degree, major or school, a prior internship, a perfect resume, or permission to try things.

Conditions

The team is based in San Francisco, and the internship is in person there.

Only US citizens or visa holders may apply.

Benefits

• Pay of $4K, $8K per month.
• Meals in the office are expensed.
• A flight stipend covers travel to San Francisco.
• Daily work alongside the founders, with code review and mentorship on everything you ship.
• A clear route to a full-time offer for interns who do strong work.