Felipe Leal
Felipe Leal

Senior Software Engineer · Sydney

I build production LLM applications.

Right now that's a voice agent that handles customer support calls end to end: it identifies the caller, finds their booking, makes changes on the call, and passes the call to a person when it should.

Focus

Around the model

Most of my work sits around the model rather than inside it

That means designing the tools and APIs an agent uses, testing its behaviour, building the data that shows how it performs, and integrating it with the systems it acts on. Before that, I spent several years on backend services and large CRM integrations.

Agent design

Tools, guardrails and hand-offs

Tool and API design for LLM agents, guardrails and hand-off behaviour.

Evaluation

Simulation-based testing

Regression suites built from simulations, so prompts and models can change safely.

Data

Pipelines and reporting

Conversation pipelines into a data warehouse, plus reporting on agent performance.

Integrations

Syncs and migrations

Two-way syncs between CRMs and internal services, and platform migrations.

Experience

Tools

Contact

Email is the best way to reach me.