NGUYEN QUANG MINH
**************@*****.***
Hanoi, Vietnam
I’m an AI Engineer working mainly on Speech to Text. Beside, I also participate in solving NLP problems
(Sentiment Analysis, Machine Translation,. . . ) and other Speech processing modules. I’m working toward the product value and looking forward to bringing my vision levitate the company development. EDUCATION
High School for Gifted Students 2012 - 2015
Department of Chemistry Graduate: Good student
FPT University 2015-2019
Software Engineering Overall score: 8.0/10.0
FPT School of Business and Technology 2019-present Master of Software Engineering Overall Score 8.7/10.0 (now) EXPERIENCE
FPT Software 05/2017 - 03/2019
Web developer
- Internal Portal Request Website:
Internal website built by .NET and AngularJS.
- Working Schedule Website/Batch application:
Project for Japanese customer. Using Java and Javascript.
- WebDAV application service on IIS:
A console application built by C#.
Vietnamese Arti cial Intelligence Solutions 03/2019 - now AI Engineer (Head of AI Engineering currently)
- Automatic Speech Recognition:
Build an ASR system by Kaldi toolkit.
Build an end-to-end ASR sytem based on Wav2Vec2 with great accuracy compare with other competitors in the market. Optimize the model to run on low-budget hardware.
- Speaker Recognition:
Develop the AI core to recognize the identity of speakers in the meeting audio using Ecapa-TDNN to extract embedding. Integrate the core in VIONE product - the intelligent meeting assistant.
- Speaker Diarization:
Stack Auditok - Xvector - Spectral clustering to build a Diarization System for VIONE product.
- Neural Machine Translation:
Preprocess data and deal with low resource problem to build the NMT for Vietnamese - English.
- Language Model management:
Manage to build n-gram language models and Vietnamese lexicon to support Kaldi-based ASR core.
- Mispronunciation Detection System:
Build an MVP product for Vietnamese mispronunciation detection.
- Data labelling:
Pioneer to build the system for labelling the data to train ASR models. Directly build the Android application to label the data.
- Others:
Mentor and support members from intern to full-time employees. Be in charge of managing the AI team, participate in making plan for the R&D department and execute them with team. PUBLICATIONS
The System for Detecting Vietnamese Mispronunciation FDSE 2021 https://doi.org/10.1007/978-***-**-****-532
Quang Minh Nguyen and Phan Duy Hung
Improving Speaker Veri cation in Noisy Environment Using DNN Classi er RIVF 2021 https://doi.org/10.1109/RIVF51545.2021.9642074
Chung Tran Quang, Quang Minh Nguyen, Phuong Pham and Truong Do Improving Vietnamese Named Entity Recognition from Speech Using Word Capitalization and Punctuation Recovery Models INTERSPEECH 2020 https://doi.org/10.48550/arXiv.2010.00198
Thai Binh Nguyen, Quang Minh Nguyen,Thi Thu Hien Nguyen, Truong Do and Chi Mai Luong VAIS ASR: Building a conversational speech recognition system using language model combination Preprint
https://doi.org/10.48550/arXiv.1910.05603
Quang Minh Nguyen, Thai Binh Nguyen, Ngoc Phuong Pham, The Loc Nguyen VAIS Hate Speech Detection System:
A Deep Learning based Approach for System Combination Preprint https://doi.org/10.48550/arXiv.1910.05608
Thai Binh Nguyen, Quang Minh Nguyen, Thu Hien Nguyen, Ngoc Phuong Pham, The Loc Nguyen, Quoc Truong Do
ACHIEVEMENTS
1. 3rd prize with Gaming Android Banking Application in VPBank Gami cation Hackathon. (2018) 2. 1st prize with Cleaning Household Service Android Application in Coding Inspiration. (2018) 3. 1st prize in the shared task of Automatic Speech Recognition at VLSP competition. (2019) LANGUAGES
IELTS: 6.0 (2015)
TOEIC: 890 (2021)
SKILLS
Technical skills:
Python, Docker, Git, Pytorch, Kaldi, Speechbrain, Bash scripting, Android Extracurriculars:
Reading, travelling, motorbike, game, gym, etc.
Languages:
Vietnamese, English.
MISCELLANOUS
I have a thing for travelling, watching movie, discovering how the magic of AI can change the world and crafting the next "big" thing in my career path.