How I work:
Statistics, Six Sigma and Machine Learning education
* These samples are based on the work I have provided for private companies. No confidential information is disclosed on this site.
Automatically generated Machine Learning models using AWS S3.
Synthetic data helps reduce the cost of data collection and data labeling. Raw synthetic data not only reduces costs, but also helps solve privacy concerns associated with sensitive real-world data. Additionally, the developer controls the distribution of synthetic data, which reduces distortion compared to real data. Including anomalies that are difficult to detect from real data can provide greater diversity. This project interacts with a financial .csv file allocated on an S3 bucket.
In-depth analysis and forecast of testing script cycle time for Facebook server products.
Iperf is a simple script test used to test network throughput, and large volumes of historical data generated by its failures and lags can be used to develop simple benchmarking of manufacturing effectivity. I used time series forecasting to design a short series of steps to approximate accurate duration of failures 24 hours ahead of time based on number of server bays, historical failures, and relative node booting lag. This method provided a 94% accuracy of expected cycle time during 6 consecutive days.
Controlled test-run for Microsoft server products. 21 hours continuous observation.
I designed a checklist divided by 10 minutes intervals to track the process and extraordinary events on the testing of servers for Microsoft server products.
Software testing can be tricky to analyze if there are human-interaction events that can potentially impact cycle time and information. I monitored over 21 hours the activity of 5 server racks to determine all possible variables in the process.
Mini statistical lab for script testing cycle time for Facebook server products.
Over 61 days I fed a matrix with average failure time by single variable to create a large historical sample of failures with color coding for quick identification of patterns. I took the largest blocks of failures and I tested their Kurtosis values to demonstrate the stability of periods of failure.
Business operation research: Identification of operations bottle neck using branch and bound method.
When walking into an unknown organization or project, a branch and bound matrix with statistical weight on interactions is a fast way to assess the critical elements in it. I used this method in my first 90 days at Wiwynn Corp. as a Lean Six Sigma Black Belt to support executive efforts in the planning for a new site in Malaysia. I successfully identified the 3 operations that held the biggest risks for the system, and kick-started the upper management focus on these areas of the company.
