Feb. 2021Ho Chi Minh City, VietnamWorking on researching / building solutions and data pipeline for several projects at Pharmacity, one of the first cutting-edge retail pharmacies in Vietnam:
- Built Recommender system by leveraging a combination of matrix factorization and graph-based approaches. Incorporated data synchronization from multiple sources to ensure comprehensive training.
- Customer segmentation supporting to coupon campaign, utilizing graph-based algorithms (Louvain and Stochastic Block Model) to detect communities of customers and analyze the network structure.
- Customer churn prevention based on transaction history, customer survey, customer service call... Utilized LightGBM's gradient boosting framework to build a high-performing model that accurately predicted customer churn.
- Pricing strategy support by collecting (crawling) product data from competitors' websites. The project has been built on top of Scrapy with Selenium integrated.
- Aggregate Delivery Tracking Data from multiple sources / vendors.
Deep Kafe Service - Technology Company LimitedData Engineer
Jun. 2018 - Oct. 2020Ho Chi Minh City, VietnamOur company mainly works on AI-related projects, such as:
- Deep Scanner: A product dedicated to interpret images of Vietnamese ID cards into digital data with high confidentiality.
- Deep Page: A product dedicated to interpret images of business document into digital data and structure them in order to search later.
- Deep Vision: A product dedicated to recognize objects or interpret videos into digital data; thus helps to make statistics & analysis reports.
- PO-Parser service that parses purchase order (PDF) automatically and import them to Odoo system of our customer (TriCoop). Because of working in a young startup, I wore many hats at work and contributed to almost all projects of the company.
- In charge of back-end stack, mainly on Python to make effortlessly scalable services; sometimes, got my hands dirty on DevOps duties.
- I built a mini framework that helps team to simplify the implementation of a system following to the Microservice Architecture and Event-Driven Architecture.
- Managed & deployed services to the cloud (using DigitalOcean and several services of Google Cloud). Underlying tech stack used by our products has a wide range:
- Flask: For exposing RESTful APIs, helps to integrate our services.
- Redis was used by almost our services for caching purposes.
- Elastic Stack: Indexing, searching, data shipping...
- We used RabbitMQ as a main messaging system that helps to route the messages, events and notifications between services.
- Apache Airflow is a great platform that we used for scheduling tasks.
- Apache Spark was using for batch processing jobs; and Airflow kicks off these jobs.
- Main databases: MongoDB, PostgreSQL.
TMA SolutionsSoftware Engineer
Feb. 2016 - May. 2017Ho Chi Minh City, VietnamProject: Alcatel-Lucent OmniVista 2500 Network Management System
Link: https://www.al-enterprise.com/en/products/network-management-security/omnivista-2500-network-management-system Omnivista 2500 NMS, helping users manage, monitor and interact with thousands of network devices easily (instead of interacting via command line), was built using Service-based architecture. As a Back-end Engineer of the project, I have developed, maintained the project with great care, collaborated to create strategic initiatives to design, code, and test solutions. I got my hand dirty on several components/features including:
- Device Abstraction Layer which used to interact with devices.
- Device Discovery service which traverses the network and collects supported devices automatically, then persists the information to MongoDB.
- VLAN feature which helps to create, manage VLAN with ease.
- Notification feature which helps to manage notification rules of devices or device types.
- Devices' usage & health reports.
Intch is a Professional Networking App for the Future of Work