Senior Data Scientist
Current- Trained and deployed models for speaker diarization and speaker verification problems.- Developed modules to assist the ASR model, including word-level confidence scoring, contextual biasing graphs, and streaming ASR.- Utilized LLM for further processing and information extraction from the textual output of the ASR model.- Led a team of three to crawl, collect, and process data for speaker diarization, speaker verification, and ASR tasks.- Built an anti-spoofing system for face eKYC processes.- Achieved 1st rank in the VLSP 2023 national challenge for Vietnamese Spoofing-Aware Speaker Verification.- Recognized as the Best Performance Employee and Team of the Year from 2021 to 2023.