Sep 2023
AWS ML Pipeline: Real-Time Data Processing Infrastructure
Data Engineering
A serverless AWS pipeline that turned six-hour batch lag into real-time business intelligence, cutting data latency to 45 seconds and saving $40K/year.
DATA & AI ENGINEER · TORONTO · AWS CERTIFIED
Phu Le
I build the cloud data platforms that power AI, production pipelines and the data-quality infrastructure that makes downstream ML trustworthy.
- RECORDS / DAY
- 3,000,000
- PIPELINE RUNTIME
- 7 min
BY THE NUMBERS
At a Glance
7+YEARS EXPERIENCE
~3MRECORDS / DAY
~80%ANALYSIS TIME CUT
$150MREVENUE INFLUENCED
FEATURED
Selected work
Jul 2023
Financial Sentiment Analysis with BERT
Machine Learning
A fine-tuned BERT model that turns a flood of financial news into a sentiment signal traders can act on, hitting 89% accuracy and cutting manual analysis time 70%.
May 2023
Customer Churn Prediction with XGBoost
Machine Learning
An XGBoost model (0.89 AUC) that flags customers about to leave while there's still time to keep them, lifting retention 8% and saving $200K/year.