Tase

A personal assistant is only as good as its memory and its judgment. Tase needs an engineer who can build the systems that turn raw model calls into a coherent assistant: one that remembers, plans, takes real actions, and gets better the more a person uses it.

About the role

You will own the AI layer: prompt architecture, model selection, agent orchestration, and the memory system that extracts and reuses what the assistant learns about each person. You will work closely with the founder and the iOS engineer to decide what the assistant can do and how it behaves. This is a systems role as much as a modeling role. The hard problems are orchestration, reliability, and cost, not just picking a model.

What you'll do

Build and refine the agent pipeline that turns a spoken or typed request into the right action. Design the memory architecture that captures facts, preferences, and context, then surfaces them at the right moment. Evaluate and integrate new models as they ship. Instrument everything so we can see where the assistant fails and fix it. Keep inference costs sustainable as we scale.

What we're looking for

• Hands-on experience building with large language models in production, not just demos.
• Strong systems engineering instincts. You think about failure modes, retries, cost, and observability.
• Comfort designing prompts and evaluation harnesses, and the honesty to measure when things break.
• Familiarity with agent patterns, tool use, and retrieval. You know where these help and where they hurt.
• Curiosity about the product, not just the tech. You want to build something people rely on.