Founding Full Stack Engineer

The Role

As our Founding Full Stack Engineer, you will work directly with the CTO and research team to build the tooling that makes our models observable, testable, and usable. You will own the entire application layer, from the database that stores model "beliefs" to the frontend that visualizes the model's reasoning trace in real-time.

• High-performance visualization tools (think D3.js or Three.js) that allow researchers to inspect model attention, memory retrieval, and reasoning paths dynamically.
• Robust pipelines that allow us to test the LTM against complex benchmarks and track performance degradation or improvement over time.
• Designing the high-concurrency APIs that will eventually serve the LTM to the world.

• Frontend:
  React, Type Script, Tailwind, Next.js. We value speed and type safety.
• Infrastructure:
  Postgres (Vector extensions), Redis, AWS/GCP, Docker.

• A Systems Thinker:
  You understand that a "full stack" isn't just a database and a div; it's understanding how latency in the backend affects the user experience in the frontend.
• Production-Grade:
  You have 5+ years of experience shipping code that doesn't break. You write clean APIs and care about database indexing.
• Curious about AI:
  You don't need to be an NLP researcher, but you should be excited to learn how embeddings work so you can visualize them better.
• High Agency:
  You spot a problem, you fix it. You don't wait for a ticket.

• Experience with WebGL, Three.js, or canvas-based rendering.
• Experience building tools for ML teams (e.g., modifying MLFlow, Weights & Biases, or custom annotation tools).