- Good Readers ETL + Build a pipeline to automatically ingest, transform, process, and store data on Airflow + Collect real-time data from the goodreads API and store it on S3 and Hadoop. Spark job reads the data and apply transformation and then transferred to the data warehouse for storage. + Build data quality and analytics modules for data stored in data warehouse + Use Airflow DAG to schedule and manage pipeline runs- Customer Data Platform + Build and install on-premises Hadoop and Bigdata systems. + Ingest and transform datasets from a variety of data sources: API, local, database...Process and analyze huge amounts of data to extract necessary information about customers with Spark, and combine with Machine Learning and Deep Learning models to analyze customers. + Participate in designing and building data lakes and data warehouses to ensure effective data storage. Combines many technologies: Cassandra, Elasticsearch, Elassandra, MongoDB. + Build a pipeline to automatically ingest, transform, process, and store data on Airflow. + API development to serve projects using data from data warehouse.- Credit Scoring + Build a process to collect data from multiple data sources and store it in a data lake. + Process, clean, and transform data to extract necessary information for research and building a credit scoring model. + Collaborate with data analysis teams to deploy credit scoring models

- Text to Speech + Building a Text to speech model for Vietnamese + Serving model Text to speech for product + Integrate Text to speech with the call center system- Text Normalize + Standardization of handwritten to spoken form for Vietnamese

- Text Classification: + Research and build a text classification model for news sites + Using PhoBERT pretrain combined with LSTM for news classification problem- Text Clustering, Text Extraction+ Extract the important keywords of the article in the text using the TextRank, TopicRanh algorithm- Labeling for Post in Social Network(Image/Video+Text)+ Research and build a news and image classification model for social networks