Shantanu Vichare

Move across the board to knock letters out of place.

Lead QA engineer at PowerSchool. I build AI agents that write, run, and repair tests.

Tests that repair themselves

I lead quality engineering for Schoology at PowerSchool. Our tests are written, run, and repaired by AI agents, and engineers review what they ship. Manual QA effort is down 80%, and Schoology has not had a P0 production incident in over four years.

  1. Fail

    A UI change breaks a locator in the assignments suite, and the Playwright run goes red.

  2. Diagnose

    A background agent pulls the logs, screenshots, and pipeline context, then checks the AI Context Graph: 10,000+ indexed tests and live OpenAPI contracts.

  3. Repair

    It writes the fix through Playwright MCP and reruns the suite. The graph grounds every step, so nothing is invented.

  4. Review

    A pull request lands for an engineer to approve and merge. Flaky-test repair time drops to near zero.

90%+ of drifted locators are fixed on the first pass, with zero hallucinated steps.Simulated run. Suite names and schematic are illustrative.

Results in production

Measured results of systems Shantanu built at PowerSchool
SystemResult
Multi-agent QA platformManual QA effort-80%
Schoology quality strategyP0 incidents in 4+ years0
API test synthesisRegression coveragewas 55%95%
Agent workflowsPrompts per taskwas 152-3
Self-healing testsDrifted locators fixed on the first pass90%+
Cloud test runnersRegression runtime-70%
Release cockpitRelease checklistwas ~4 hours~15 MIN
DevSecOps blueprintVulnerabilities caught before production30+

Toolkit

Agents and LLMs

  • Claude Code
  • Cursor
  • GitHub Copilot agents
  • LangChain
  • RAG
  • Multi-agent orchestration
  • Claude, OpenAI, Gemini
  • Lovable

MCP servers

  • Playwright MCP
  • Browser MCP
  • Jira MCP
  • AI Context Graphs

Test automation

  • Playwright
  • WebdriverIO
  • Selenium
  • Robot Framework
  • Codeception
  • API testing
  • Visual regression
  • BDD

Workflow and CI/CD

  • n8n
  • GitHub Actions
  • GitLab CI
  • Bamboo
  • Containerized runners
  • Ephemeral environments

Security and load

  • OWASP ZAP
  • Burp Suite
  • SAST and DAST
  • Dependency scanning
  • JMeter
  • LoadRunner

Languages

  • Python
  • TypeScript
  • JavaScript
  • PHP
  • MERN stack

Eleven years, three companies

  1. Lead Quality Assurance EngineerPowerSchool

    2020 - present, Bengaluru, remote

    • Promoted into AI quality engineering leadership in 2023 after three years leading Core QA. Own AI test tooling strategy for Schoology and mentor 20+ engineers across 6 Agile squads.
    • Architected a multi-agent test platform: GitHub Copilot background agents orchestrating Playwright, Browser, and Jira MCPs, with Claude, OpenAI, and Gemini models authoring, running, and triaging tests.
    • Built the AI Context Graph that anchors agents to live OpenAPI contracts and canonical suites, so tests repair drifted locators with 90%+ first-pass accuracy.
    • Replaced manual Postman work with agent-generated Playwright API tests from OpenAPI specs, lifting regression coverage from 55% to 95%.
    • Scaled Cursor and Copilot agent workflows across the tribe: prompts per task down from 15 to 2-3, and 4+ engineer-weeks reclaimed per release cycle.
    • Ran containerized Playwright and WebdriverIO runners on ephemeral cloud environments, cutting regression runtime by 70% and unlocking same-day production deploys.
    • Embedded OWASP ZAP, Burp Suite, dependency audits, and policy gates in Bamboo, GitLab, and GitHub Actions, blocking vulnerable code at merge.

  2. Software Development Engineer in TestFireEye (Mandiant)

    2018 - 2020, Bengaluru

    • Built UI and security test frameworks from scratch for Linux, Windows, and macOS endpoint products, cutting escaped defects ~35% release over release.
    • Wrote Python and Nosetests suites that validated detection and prevention across 8+ security releases, shortening regression cycles ~40%.
    • Triaged complex security bugs with developers and product managers across 5+ distributed teams.
    • Spread Python, Selenium, and Robot Framework practice to 15+ incoming engineers.

  3. Associate Threat Research EngineerQuick Heal Technologies

    2015 - 2018, Pune

    • Built internal tooling that replaced manual malware analysis, reclaiming ~12 analyst-hours a week.
    • Led customer escalations and root-cause analysis across 50+ security incidents, feeding fixes into the detection roadmap.
    • Proposed 2 threat-detection capabilities that shipped in Quick Heal products used by millions.

Education

  • 2025 - 2027

    MSc, Machine Learning and Artificial Intelligence

    Liverpool John Moores University, England

    In progress

  • 2025 - 2026

    Post Graduate Diploma, AI and ML

    IIIT Bangalore

  • 2012 - 2015

    Bachelor of Engineering, Information Technology

    Shah and Anchor Kutchhi Engineering College, Mumbai

Certifications

  • 2023CopilotXcelerate: AI-Powered Python for Data ScienceUpgrad
  • 2019Certified Ethical Hacker (CEH)Simplilearn
  • 2016Oracle Certified Expert, Database SQLOracle

Open to conversations about AI quality engineering, multi-agent systems, and automation architecture.

Select the board to copy the address.

Based in Bengaluru, India.