Built database backend developer tools in Java/Spring Boot and C++ that power data engineers to ingest data to the database, data scientist to design and manage metadata in the database, and develop and run queries against the database to explore data and their relationship; promoted customer privacy in collecting debugging info for technical support team with Regular Expression masking/obfuscating user sensitive info; stabilized products via robust test through unit test and e2e test with JUNIT and shell, object-oriented programming under Linux environment.

* Enhanced database for OpenCypher to expand customers in the graph data area, with support for common string and math functions, and statistics functions via lazy evaluation to avoid overhead computation, in C++.

* Unified database interaction with different programming languages, and refactored and migrated authentication APIs/command-line to REST API with Spring Boot in Java, by designing JSON payload, url, method, header, request body, response body and status code.

* Enhanced security for user-privilege operation under database developer command-line interface (cli) by masking passwords with Java ConsoleReader, and recycling locks in designed order regarding concurrency issues.

* Concealed customer sensitive info in logs with Java Regular Expression.

* Developed a debugging info collection tool in Go for customer’s yaml config files with Go Interface, masking sensitive info (schema) via Go Regular Expression as well as unit testing with Go package testing.

* Improved HeapAcum accumulator that helps users store and sort tuple data by unifying comparator function in Java and C++.

* Designed canonical tests to cover function parameters with a Python driver tool.

* Resolved urgent database crashing bugs regarding semantic checks, query optimization, code generation, and type inference.

Deep learning Natural Lauange Processing (NLP) models using Pytorch in Python for identifying Android malware (US National Science Foundation, NSF project) under cloud-environment (AWS) in collaboration with security researchers.

* Developed a fine-tuned BERT NLP model, and utilized AWS Boto3 library to dynamically load data into instance, with a customized lazy tokenization to avoid overhead loading.

* Evaluated guardraisl against hazards with GIS on low-volume rural roads (Kansas Department of Transportation, KDOT project), with a publication approved by KDOT instructing local engineers regarding road-hazard security assessment.

* Improved real-time traffic monitoring by calibrating with Bland–Altman plot (KDOT project) for TransCore.