SDET  ·  QA Automation  ·  AI Agent Testing  ·  Bengaluru

KARAN SONI

Shipping quality at every sprint.
Automating what slows teams down.

Work with Me

I build reliable automation frameworks that give engineering teams confidence to ship fast. From Selenium to AI agent QA — I make quality a feature, not an afterthought.

85%
Regression Coverage
70%
Manual Effort Reduced
300+
AI Agent Scenarios
92%
NLU Accuracy

About

Quality is a system,
not a checklist.

SDET & QA Automation Engineer with 1.5+ years building test infrastructure across web, API, mobile, and AI-agent platforms — Selenium, Appium, Rest Assured, BDD Cucumber.

At Superfone in Bengaluru I lead quality engineering in Agile sprints. I also pioneered AI Agent QA for the company — designing test suites that validate NLU intent recognition, entity extraction, and multi-turn conversation flows for chat and voice agents.

Regression coverage moved from patchy to 85%+. Manual testing effort dropped by 70%. Release cycle time cut by 30%.

Bengaluru, India Agile / Scrum Shift-Left Testing AI Agent QA BDD / Cucumber Open to Opportunities
0%
Regression Coverage
0%
Manual Effort Cut
0%
Faster Releases
0%
Faster Resolution
Selenium WebDriverAppiumWebDriverIOCypressRest AssuredPostmanJavaJavaScriptTestNGBDD CucumberPage Object ModelMavenGit & GitHub Selenium WebDriverAppiumWebDriverIOCypressRest AssuredPostmanJavaJavaScriptTestNGBDD CucumberPage Object ModelMavenGit & GitHub
Chat Agent QAVoice Agent QANLU ValidationSarvam AICartesiaGeminiJIRAAgile / ScrumSDLC / STLCShift-LeftAndroid StudioDefect Management Chat Agent QAVoice Agent QANLU ValidationSarvam AICartesiaGeminiJIRAAgile / ScrumSDLC / STLCShift-LeftAndroid StudioDefect Management

Experience

Where I've shipped quality.

QA Associate — Automation & AI Agent Testing Superfone Sep 2024 – Present · Bengaluru
  • Built Selenium + Java + TestNG framework — regression coverage to 85%+, release cycles cut by 30%.
  • Automated 80% of manual regression — weekly runs dropped from ~8 hours to under 2.5 hours per sprint.
  • Led 300+ Chat & Voice Agent test scenarios (intent, entity, fallback, multi-turn) — achieved 92% NLU accuracy post-retraining.
  • Owned staging readiness gate before every production release; managed utterance dataset retraining cycles.
  • Reduced mean time-to-resolve defects by 25% through structured JIRA tracking across web, API, and AI workstreams.
Graduate Trainee — QA & Infrastructure Cognizant Feb 2024 – Jun 2024
  • Delivered QA across 3 internal client projects under STLC — zero critical defects reached production.
  • Improved defect documentation quality via CIS workshops — faster cross-team handoffs and resolution.

Projects

Selected work.

01 — Framework

Selenium POM + BDD Framework

Page Object Model automation with BDD Cucumber — non-technical stakeholders co-author Gherkin scenarios. Built for scale, reuse, and long-term maintainability across sprint cycles.

85%Code Reusability
60%Less Maintenance
JavaSeleniumTestNGCucumberBDDMaven
GitHub

02 — Mobile

Appium Mobile Automation Suite

150+ Android test cases across login, navigation, and transactional flows. Replaced manual device testing — confident weekly releases with zero regression debt.

150+Test Cases
65%Faster Regression
AppiumJavaTestNGAndroidAndroid Studio
GitHub

Background

Education & Certifications

Masai School
Diploma in Software Testing & Automation
Jun 2024 – Sep 2024GPA 8.5 / 10
Dr. Babasaheb Ambedkar Marathwada University
B.S. Computer Science
Feb 2020 – Jul 2023GPA 8.2 / 10

Certifications

QA
Software Testing in Automation
Masai School  ·  View Certificate ↗
AI
Career Essentials in Generative AI
Microsoft & LinkedIn  ·  View Certificate ↗

Contact

Let's build something
that doesn't break.

Open to SDET, QA Automation, and AI Agent testing roles. Based in Bengaluru — available for full-time positions.

bussinesskaran6@gmail.com
+91 94033 43224 LinkedIn GitHub