b0noi / aboutme Goto Github PK

email: [email protected]

• Facebook

• Linkedin

I am an expert in performance testing.

I create performance tests and build processes for finding performance issues. I help companies to understand why their programs work slowly. I teach people to look for and prevent such kind of issues.

I am looking for problems not only in programs but also in processes. I define the requirements, and I create test scripts based on those requirements for various test types and load profiles. I setup a pipeline of scenarios in a continuous integration system and setup a system for automatic analysis and anomaly search using machine learning. I keep human interaction with tests to a minimum and provide developers with the fastest possible feedback after their commits.

Companies contact me when they need to understand:

• why the program is slow ;
• when it is necessary to build a process for detecting performance issues from scratch;
• when it needs to teach the system automatically find issues during the development process.

I can work with many performances, monitoring, and profiling tools. I can find errors in desktop applications, web applications, and databases. I can independently analyze large volumes of logs and find hidden correlations between several data sources. I can use machine learning to find problems of balancing, memory leaks, and predict future resource consumption by servers.

I’m not a wizard, and I can’t find errors and bottlenecks just by looking at the application. Load testing is scrupulous research that required long preparation and gives accurate results. However, we can always find a compromise by sacrificing accuracy for the reason of speed (I prefer not to do so).

My mission is to create one button - which can say why the application is slow

• Anomaly detection in performance testing (static)
• Anomaly detection for monitoring system (dynamic via Plumber)
• Add machine learning in performance testing
• Add automatic results analysis of performance testing
• Generate recommendation system for performance testing engineer for a faster gathering of performance testing results
• Create interactive shiny-dashboard for performance testing results
• Create a robust regression model for faster gathering of performance testing results
• Create CI performance framework in a company
• Invent performance testing pipeline in a company
• Structured performance methodology in a company

• NLP: QA-fest review analysis
• Regression: Predict performance behavior on high load based on low load.
• Forecasting: Create expected requirements for performance testing based on historical vales
• Other: detect anomalies in time series for bug and route cause detection on performance testing

• Develop methodology of performance testing management: "From chaos to application performance management"
• Develop a methodology for faster gathering performance testing results using machine learning
• Develop strategy (algorithm) about getting information about changes in application performance in one performance run
• Develop an approach for predicting expected result without defined requirements (based on historical results and ARIMA forecasting)

• Educate students in National aviation university for performance testing
• Mentoring and teaching junior performance test engineers

• QA FEST - How to test app which is using machine learning - (RUS) Как тестровать приложения которые используют машинное обучение
• QA Fest - How save time on performance testing report analysis - (RUS) Как сэкономить время на анализе отчетов о нагрузочном тестировании
• QA Fest - ML usage methodology in performance testing - (RUS) Методология использоания машинного обучения в нагрузочном тестировании