Applied Myth

Areas

AI systems and agents

AI gives us a way into work we could not easily attempt before. We use it to build, learn, and connect tools. In selected products, we also give it a specific task for the person using the app.

Give it something to work with

A model needs more than a request. Saved information, project decisions, and useful references change what it can do. Emberel puts part of that question into practice: how can a quick capture become something useful later?

Connect the right tools

We try local and hosted models, agents, and supervised tool access. Some jobs need a short conversation. Others need the system to read information or carry out a sequence of steps. The work tells us which setup is worth trying.

Check what happened

A convincing answer can still be wrong. We inspect the output, try the code, or follow the task through to its result. We keep track of what worked and what still needs a person to decide.

Related work