Quyet Do

Work on Vision Language Models. Mentored by Viet Lai.

Support the software engineering team, implement a chatbot for psychological therapy.

I work on the whole pipeline, given an entity's name then scrape data on-the-fly and summarize core information. I mainly focus on the core (NLP) engine of the solution, including Data Crawling, Information Extraction. Sentiment Analysis and Summarization are other two components that I implement for the engine, but with less effort. Please visit my blog for more details https://dovanquyet.github.io/posts/en/summer-internship-2022

I started to professionally build my career in NLP here. I help a PhD student in training and testing models, as well as conduct my own experiment with his guidance. Also, I self-read research papers to gain more knowledge and propose some ideas. Python, PyTorch, git, tmux, Unix CLI are used to work on the projects. I'm also familiar with TPU training. Update on June 2022: We submitted our research paper to EMNLP2022. Let's wait!

In the first two months, I studied Computer Vision and assisted on-going projects of the team. Warming up with basic Machine Learning, I step-by-step dived deep into prevalent algorithms and neural networks in CV. Python was used for the whole process, with support of common frameworks in CV like OpenCV, PyTorch, TensorFlow, ... At the third month, I switched to Natural Language Processing as a beginner. Starting with some specialized knowledge in linguistics, I learned how to manipulate words and texts in vector representation. After that, I studied many famous models as RNN, LSTM, Attention mechanism, Transformers, BERT, etc. and the history behind those invention. I was the key person of an internal project in Machine Translation. I took part in all steps in MLOps cycle, say, I looked for data sources, processed it and trained/tested model, then deployed to web application. The project is considered as successful, and continues to improve.

I helped my senior, who is a PhD candidate, in technical aspect: - Crawled and processed raw data of chemical reactions (~ 100GB of data) - Applied Machine Learning and Reinforcement Learning to tasks in Reaction Graphs - Utilized various OCR tools and proposed ad-hoc algorithms to extract data from research papers

• Did pre-class preparation, graded and corrected student's homework, supported students during the classes, sometimes taught the classes • Did monthly evaluation and organized/prepared bimonthly tests for students • Helped students who need to catch up the regular classes for enrollment

PhD student at CS@VT. Supervised by Prof. Tu Vu (https://tuvllms.github.io/) Aspire to develop AI that evolves through efficient self-learning, similar to the development of human intelligence. Research interests: multi-instruction following evaluation, creative reasoning, long-context task solving, self-learning agents with memory.

I continue my study at HKUST as an MPhil student under the supervision of Prof. Yangqiu Song. From Nov 2022 to Jun 2023, I collaborated with Prof. Pascale Fung. Currently, I inherit several research interests of my mentor, thus focus on Conceptualization in Commonsense Reasoning. In the future, with the emergence of LLMs, my research lies on the design of LLMs to enable cognitive functionality in the models.

Double majors in Data Science and Pure Math (Advanced) Noticeable courses (with letter grades): - Honors Design and Analysis of Algorithms (A+) - Honors Object-Oriented Programming and Data Structures with C++ (A+) - Capstone Project (A+) - Big Data Mining and Processing, focusing on NLP (A) - Cloud Computing (A+) Besides, some interesting courses I have taken are - Humanity: Love, Death, and Human Nature (A) - Intro. to Electronic Music Composition (A) - Law in Society and Business (A)

Student of "Cử nhân Tài năng" (Talented Bachelor) class in Hanoi University of Science. During this time, I applied to HKUST and got admitted.