Integration Services (8 months)- Integrate features between 2 systems via APIs.Responsibilities- Involve taking customer requirements and understanding their problems, offering the most optimal solution based on the available features of the product (or having to update more).- Maintain and update features with the solution agreed to APIs to integrate with customer systems.Achievements- Successfully integrated Oracle and SAP of customers such as Mega Market, Circle K, Ân Nam, and The Coffee House.Purchase Order Processing Service (8 months)- Construct a service that receives P.O. (purchase order) emails, reads, and extracts P.O. (purchase order) information from PDF files to create S.O. (sales order) on the clients' system.Responsibilities- In charge of building the service from scratch: design database, APIs for integrating client's system, source code organization, Serverless configuration, code optimization, deployment, and testing.- Writing unit test coverage is around 75%.- Cover the entire P.O. template of the client's partner (around 30 samples).Achievements- Service is still maintained and active up to the present time.- The system can handle Purchase Orders files with up to 1000 pages.Invoice Information Extraction using Deep Learning (6 months)- Build A.I. models to extract the data required from VAT invoices.Responsibilities- Collect, filter, and process data to ensure that models are as accurate as possible.- Investigate deep-learning methods for locating and extracting data from invoices such as YOLO and Faster RCNN for text localization tasks. CRNN, VietOCR for a text recognition task.- To achieve the project's deadline and ensure that each invoice template is accurate.Achievements- Allowed to participate directly in the deployment of models to the production environment.- Optimize the predicting time of A.I. models.