← All work

Industry work

Voice-agent simulation and evaluation

A simulation harness for realistic caller personas, automated scoring and behavioral regression tests.

2025–2026Role period

The work

Built a voice-agent evaluation platform with a simulation harness for configurable caller personas and realistic phone-call scenarios, automated quality and safety scoring, and regression testing before deployment.

Context

Voice AI, evaluation, and enterprise systems.

Scope

Forward Deployed AI Scientist at BCG; primary engineer and technical owner of the voice-agent simulation and evaluation platform.

Confidentiality

Client source code and implementation details are not publicly shared.