- Developed scalable microservices to pull external data and serve analytical results to business operations.
- Collaborated with Data Scientists/Engineers, developing ML models to make insurance smarter.
Hung Do
About
Hung is a full-stack data scientist with expertise in both math statistics and agile software development. She has processed financial valuations of multi-million dollars for agricultural corporates, analyzed clinical data for major hospitals in Australia, and developed machine learning models for the EuroLeague - Europe’s premier basketball competition. She brings a wealth of international experience to her teams, having worked and studied in Singapore, France, Germany, Switzerland, and Australia. She completed her Ph.D. at UNSW Sydney with a Dean’s Award Nomination for the Top 10% of Theses.
Employment
- Discovered a good business idea through market research, customer survey, and front-end prototyping.
- Researched suitable tech stacks and break the MVP down into small doable projects.
- Recruited highly talented students from USA and Vietnam, identify their strengths and weakness, standardized training materials, and train them to prioritize their own tasks and work independently.
- Tools used by team: TypeScript, Python, GitLab, AWS EC2, WordPress SEO, Google Optimize.
- Results: Free Extension published on Chrome Web Store, marketing campaign launched.
- Product: A decision support system for financial institutions lending to agricultural companies.
- Developed APIs to Extract-Transform-Load (ETL) data and cross-check with government databases.
- Developed valuation and machine learning (forecasting…) models, updating the credit continuously.
- Strategized software migration, data security, DevOps plans, and managing junior developers.
- Scaled up the software, deploy microservices to the distributed clusters, and do network engineering.
- Prototype Tools: API (SOAP/REST), Python (SQLAlchemy, Facebook’s Prophet…), AWS (EC2, RDS, ECS-Docker, Route53, CloudFront, Load Balancer...), JavaScript (Node-Express), Linux, GitLab CD.
- Scale-up Tools: Django, Microservices Architecture, Terraform, Multi-Cloud Strategy.
- Results: Software fully automated, processing valuations of multimillion-dollar every week.
- Aim: Work with two major hospitals in Sydney. Migrate data from the legacy databases to AWS.
- Built an analytics platform for Data Science Research and Real-time Clinical Decision Support.
- In a team of four, strategized system architecture and prototyping.
- Developed the workflow and scalable software for the data pipeline, from Extract-Transform-Load (ETL) to Data Normalization and Data Quality Control (via a Dashboard).
- Tools: Amazon Workspace, Python (SqlAlchemy…), R (SqlRender, Shiny), PostgreSQL, OHDSI medical open-source tools, CERNER hospital information system.
- Results: Delivered the data pipeline, discussed tests and use cases with the hospitals
- Product: An AI platform that provides real-time market information to crypto traders, helping them maximize profits while minimizing risks.
- Managed a core team of 6 people, including external consultants and in-house developers.
- Was responsible for the development of the ICO web app for the token sale.
- Set up the continuous deployment pipeline from GitLab to AWS, connect the ICO payments to the blockchain and analyze sale data.
- Wrote the user stories (requirement document) for the website, distribute the workload and coordinate the progress between team members. Trained a student with progressive tasks.
- Tools: JavaScript, Python, AWS (Serverless, Lambda, S3, Cognito, DynamoDB…), GitLab CI/CD.
- Result: The ICO website collected around 100,000 Euros during the first few weeks of opening.
- Aim: Develop IoTA models for EuroLeague - the premium basketball competition in Europe.
- Automatically analyzed videos to create highlight snippets in real-time.
- Contributed to all stages: brainstorming, feasibility study, modelling, productization, and validation.
- Analyzed time-series data from IoT sensors (accelerometers) with Machine Learning and Signal Processing algorithms. Continuously integrate new models into the Big Data platform.
- Automated data cleansing, and wrote a webpage in PHP/ MySQL for mass data annotation.
- Tools: Python(keras, sklearn, xgboost…), Linux, Elastic Search, Atlassian, PHP/MySQL.
- Results: Mobile app “HEED EuroLeague” on iOS and Android went live during basketball season.
- Aim: Quantum Cryptography for Satellite communication in Low Earth Orbit.
- Worked with tight deadlines and industrial standards (Technology Readiness Level 8/9).
- Analyzed data from space-endurance tests. Programmed embedded microcontroller.
- Developed Software (Tcl/Tk in Linux) for real-time visualization of the signal.
- Tools: Cypress Programmable System-on-Chip, Tcl/Tk, Linux, Altium Designer, Autodesk Inventor.
- End-Result: Payload survived NASA Antares explosion 2014, successfully re-launched 2015.
