Research Scientist · Google DeepMind · New York City

Joanna Hong

Speech, audio, and multimodal AI.

I'm a research scientist on the Gemini Audio team at Google DeepMind, where I work on speech and audio understanding for Gemini, from core model capabilities to safety alignment for audio-in and audio-out launches.

Before that I helped build multilingual LLMs at Trillion Labs, interned at Meta Reality Labs, and received my Ph.D. from KAIST with Prof. Yong Man Ro. My thesis on human speech understanding through multimodal representation learning received the Outstanding Dissertation Award.

Portrait of Joanna Hong
News
ProjectsView all

Gemini 3.5 Transcribe

Google AI Studio · 2026

Intelligent transcription with Gemini, available to developers in Google AI Studio.

Speak to Window

Gemini app for macOS · 2026

Natural voice input in the Gemini desktop app for macOS, bringing speech and multimodal AI into people’s everyday workflows.

Rambler in Gboard

Gboard on Pixel 11 · 2026

Gemini-powered voice input built for the way people actually talk: it transcribes your point cleanly and concisely, even through “um”s and “ah”s.

Gemini 2.5 Pro & Flash TTS

Google Cloud Vertex AI · 2025

Gemini text-to-speech models made generally available to developers and businesses on Vertex AI. My first launch at Google DeepMind.

Selected publicationsView all