I am a freelancer in uTest. I focus on functional testing in mobile end and web testing.

Madz Adventure tries to use the most common and concise programming model to build a rapid developing platform.1. Developing/Testing Lifecycle Framework on github.com.2. Build an ETL product migrating MS Access data to MySQL and integrate to another data service with RESTful services with Java7 and JavaEE7.Major Skills:Test driven development, Core Java, JPA, EJB, RESTful, JUnitMajor Tools:Eclipse, Glassfish V3, MySQL 5.6

I had worked for 2 years and 8 months as a Microsoft SQL Engine Onsite Team member.Mainly responsible for the investigation of SQL Server module function testing and maintenance of test cases, found 180 product defects and 268 test defects, repaired 144 test cases, improved the robustness of the product and test cases; the second is responsible for writing the Plan Cache test case under the framework of TestShell, added 25 new test cases, obtained the customer acclaim and the company's most innovative Award; finally responsible for technical support of the new testing framework for our group.

Mainly responsible for Microsoft speech recognition engine training acoustic model and language model, compiling acoustic model in the windows environment, and testing and optimizing model. Has been responsible for training model of Mexico Spanish, Norway Bakemo language. The main technology: Perl, Product Studio, HTK, BAT, programmingConcrete work is as follows: In summary, there are 6 stages for a model training: preparing training and testing data, training model, compiling model, testing model, analyzing results, improving model. For preparing training and testing data stage, we have to check the quality of each language corpus in order to ensure the performance of recognition. Preprocessing the corpus marker by Perl, then the data can be used in training and testing stages. Extract feature and code of Wav file, and compile the wav files into MFC files. Training a Letter-to-sound rules to generate the pronunciation of the words. Generate training dictionary, the Vendor dictionary, test scripts and test grammar. For the training model stage, under the framework of Autotrain, with the training data for acoustic model training. For compiling model stage, in the windows environment, compile model.For testing stage, use test data to test the compiled model. For analyzing results stage, evaluate and analyze the test results.For improving stage, optimize configurations to improve overall performance of the model.