Bull's blog Bull's blog
Resume
  • MBTI Test
  • SBTI Test
  • AI Persona Test
  • Tools Home
  • Testing Toolbox
  • Test File Downloads
  • Image Test Files
  • Audio Test Files
  • Video Test Files
  • Document Test Files
  • Pinyin Dictation
  • English Word Daily
  • Paper Games
  • AI Podcast Generator
  • AI Music Generator
  • Morning Checklist
  • Work Notes
  • Categories
  • Tags
  • Archives
  • zh-CN
  • en-US
  • ja-JP

Bull

Resume
  • MBTI Test
  • SBTI Test
  • AI Persona Test
  • Tools Home
  • Testing Toolbox
  • Test File Downloads
  • Image Test Files
  • Audio Test Files
  • Video Test Files
  • Document Test Files
  • Pinyin Dictation
  • English Word Daily
  • Paper Games
  • AI Podcast Generator
  • AI Music Generator
  • Morning Checklist
  • Work Notes
  • Categories
  • Tags
  • Archives
  • zh-CN
  • en-US
  • ja-JP

Wang Yang

Senior QA Engineer (AI Testing) · Quality Engineering & SDET

✉️ cowboy231bb@gmail.com 📱 13520882907

Quality engineering experience across AI, fintech, data platforms, enterprise systems, and consumer finance, recently extended through a part-time full-stack AI Coding practice for clinical specimen repository automation.

📄 Download PDF
1
Traditional QA Foundation
Solid hands-on experience in automation, complex business modeling, data governance, and performance testing
2
AI Transformation + Coding
AI-driven test design, Agent execution, Human-in-the-loop, full-stack AI Coding, and complex integration validation
3
Execution + Communication
Solves problems on the ground while also driving knowledge sharing, training, and team collaboration

Work Experience

Reverse Chronological
SenseTime AI Research Institute
Dec 2025 – Present
AI QA Engineer / AI-Driven Test Engineer
  • AI-driven test design: Digital test assets + Agent enable Case design and smoke execution in 30–60 min
  • AI toolchain deployment: Structured interface test tasks handed to Claude Code / Codex / OpenClaw; ~85% of interface functional tests are executed with AI tools
  • Human-in-the-loop: Preserved manual review for input quality, result credibility, and log traceability
  • Role evolution: Continuously optimizing test document library and case templates, pushing testers toward AI test flow designers
  • AI performance testing: Covers HTTP interfaces and WebSocket real-time voice scenarios, tracking first-packet latency, real-time factor, capacity, and GPU load
  • Load testing: Designed Locust-based load plans, analyzing throughput, latency, P95, and capacity ceilings across single-replica and six-replica concurrency gradients
Ubiquant Private Fund Management System
Aug 2023 – Jun 2024
Senior QA Engineer
  • Business modeling: Decomposed complex fund operations by entity/field/rule relationships, building test models that go beyond page-level verification
  • Mock data system: Designed a complete Mock data framework to work around confidential data restrictions, improving executability and reusability of complex scenarios
  • Script refactoring: Parameterized Robot Framework scripts and restructured in BDD style, improving maintainability
  • Data governance: Drove unified field definitions and value standards across upstream and downstream systems, reducing integration and issue-location communication costs
Xiaomi CRM / RMS / Dynamics 365
Aug 2022 – Aug 2023
Senior QA Engineer
  • Overseas business testing: Covered product distribution, store management, sales reports, and staffing, adapting to multi-country variance paths (tax filing, export, duty-free thresholds)
  • Low-code platform testing: Rapidly grasped Dynamics 365 entity/field/reference design patterns and translated them into executable test strategies
  • Automation from 0 to 1: Built interface automation capabilities and delivered UI automation Demo/PoC
  • Long-chain clarification: Calibrated test case directions against requirements written based on assumptions
Tianyancha OpenAPI / Pro Edition
Feb 2021 – Aug 2022
Senior QA Development Engineer / QA Team Lead
  • Data product testing: For multi-entity data products, clarified test boundaries and anomaly categories to meet clients' high sensitivity to data accuracy
  • Patrol & alerting: Built automated patrol + data prep + Feishu alerting (with fault tolerance) suite, improving anomaly localization efficiency
  • Anomaly governance: Developed a data-product testing approach centered on "anomaly root-cause classification + data quality governance"
MaTongXue (Code Classmates)
Jan 2020 – Mar 2021
Automation Testing Instructor (Full-time)
  • Course delivery: During a 15-month tenure, served 300+ students and completed a 17-week curriculum design and instruction cycle
  • Curriculum iteration: 3 revisions, introduced pytest / httprunner / airtest into content
  • Content production: Produced RF, Airtest, Git recorded courses for platform distribution; Douyin live streaming reached 300+ concurrent viewers
MaShang Consumer Finance
Nov 2016 – Aug 2019
QA Engineer → QA Team Lead
  • High-frequency delivery: Supported an average of 2 releases/week across Official Account, Mini Program, and H5 channels
  • Team management: Established test demand pool, AB-role mechanism, and resource coordination to improve test throughput and collaboration stability
  • Risk stratification: Under high-frequency change and limited resources, completed priority management to safeguard critical module quality

Featured Projects

AI Testing Engineering

Case A: AI Test Assets & Agent Execution

In Progress
Problem Experimental AI projects with extreme short cycles, no API docs, traditional testing breaks down
Action Structured test intent into Agent-readable assets, built Human-in-the-loop gates, and captured reusable prompts and constraints after each retrospective
Result New AI demand response compressed from one day to 2-3 hours; team startup cost for similar demands significantly reduced
PythonClaude CodeWebSocketJenkinsQuality GatesTest Assets

Case B: Specimen Repository Automation — End-to-End Delivery

Part-time Practice · Jul 2026 - Present
Problem A legacy business system, local bindings, three device flows, and manual handling all changed operational truth while requirements and protocols evolved on site
Action Used AI Coding to deliver Vue / FastAPI / SQLite, adapters, Mock environments, automated regression, device terminal-state handling, and controlled recovery
Evidence Scanner field records; Aug 24, 2026 offline snapshot: backend 1051 passed / 7 skipped, frontend 92 passed with a successful production build
AI CodingPythonVue 3FastAPISQLiteRabbitMQLayered Validation

Case C: ASR Transcription Benchmark

Completed
Problem ASR accuracy relied on manual listening, standards varied by person, no objective regression after model upgrades
Action Reproduced Hugging Face framework, built C/J/E dataset, established cleaning rules, introduced LLM for deviation grading
Result Objective accuracy report available on model release day; manual review compressed from full listening to priority checking by deviation level
PythonASRHugging FaceLLMData CleaningBenchmark

Skills

Current Stack
PythonPytestPlaywrightLLM Prompt EngineeringAI-Assisted TestingLocustJenkinsWebSocket TestingGitLab CI
Previous Experience
JavaTestNGRobot FrameworkSeleniumAppiumJMeterMySQLRedisDockerAutomated Patrol

Leave me a message

Questions about my experience, projects, or collaboration are welcome.

No messages yet. Be the first to leave one.

cowboy231bb@gmail.com · 13520882907
© 2026 Wang Yang · Updated 2026/08/25
上次更新: 2026/08/25, 03:33:50
最近更新
01
asr-translation-accuracy-benchmark
08-25
02
resume-fde
08-25
03
case-d-specimen-automation
08-25
更多文章>
Theme by Vdoing | Copyright © 2018-2026 Evan Xu | MIT License
  • 跟随系统
  • 浅色模式
  • 深色模式
  • 阅读模式