AK.

Mission: dependable software

Ankit
Kalluraya.

Software QA Engineer & Test Automation Specialist

From rigorous software testing to applied AI experiments. I build, challenge and refine systems, with evidence at every stage.

01 / AI quality02 / Applied AI03 / Automation
AK / MISSION CONTROLSYSTEMS READY
MISSION 01 / 03PHASE PRE-FLIGHT

Define the risk. Build a testable system before launch.

01 / SELECTED WORK

Recent missions. Practical evidence.

Explore recent workspace and simulation projects, then inspect the AI quality tools and their documented scope.

Recent work · Local-first workspace

B-Ram

A browser workspace for notes, PDFs, bookmarks and code. Local IndexedDB storage, full-text search, source excerpts and backup tools make collected information easier to revisit.

Optional AI and embedding features depend on configuration. Personal workspace prototype, not a claim of production deployment or customer adoption.

IndexedDBSearchLocal-first

Recent work · Interactive 3D prototype

Orbit Desk Companion

A Three.js desk-companion concept with simulated sensors and voice interactions. Its Assembly Arcade includes five build missions, placement and wiring checks, time playback and downloadable MP4s.

A browser simulation for exploring assembly and interaction. Not a built robot, hardware validation or safety certification.

Three.jsSimulationInteractive learning

Flagship · AI QA

ReleaseRadar

A focused regression-testing prototype that exposes gaps in a passing checkout test suite. Freeze assertions, challenge three deliberately seeded bugs, review proposed additions, and rerun unchanged assertions against the corrected build.

Prepared and imported cases run in the hosted demo. Live test generation requires a configured local server. Scope: one checkout requirement, not release certification.

JavaScriptOpenAIPlaywrightGitHub Actions

Open-source adaptation · Applied AI

Jev Flight Agent

An adaptation of Browser Use’s Jev Ultrafast browser agent for natural-language flight searches. Jev selects typed operations and observed DOM targets; a separate language model supplies field text, and Playwright executes the actions. The loop checks stale page state and bounds execution.

Core agent, demonstrations and benchmarks by Gregor Žunič / Browser Use. My contributions: documentation, repository setup and UTF-8 compatibility fixes for Windows. Local search prototype requiring API keys and Chrome, not a booking service.

PythonJev / TypeSafe AIPlaywright

Supporting · AI evaluation

Assay

Calibrates an LLM judge against human-labeled examples using Cohen's kappa. The default gate blocks uncalibrated scoring; reports retain calibration provenance and track model-call budgets.

PythonOllamaLLM evaluation

Supporting · Independent AI

Jev Ticket Router

A FastAPI prototype that classifies support tickets by team, urgency, and customer sentiment in a batched Jev call. Department confidence below 0.6 sends the ticket to human review.

Requires a TypeSafe API key. Prototype scope; no production-service claim.

PythonFastAPIJev

Supporting · AI QA

LLMTestKit

Local-model evaluation and red-team tooling with deterministic checks, a correctness judge, prompt-injection and jailbreak cases, and HTML reports with score provenance.

Early-stage toolkit. RAG, groundedness, and broader agent evaluation remain roadmap items.

PythonOllamaRed teaming

Supporting · Test automation

Playwright Test Automation Framework

Python and Pytest UI automation for SauceDemo, organized with page objects and JSON test data. Includes parallel test execution and a GitHub Actions workflow.

PythonPytestPlaywright

Explore all public repositories ↗

🎓 Education

The University of Melbourne, Melbourne, Australia
Master of Engineering  |  2017 – 2021

Work Experience

Independent AI Quality Consultant & Founder

Independent · Full-time
Aug 2025 – Present
Australia · Remote

  • Built a multi-agent adversarial framework (Doer/Adversary/Judge) to detect prompt injections and hallucinations, validating outputs from GPT-4, Claude 3.5 Sonnet, and Llama 3.
  • Engineered an AI Agent Validation framework using LLM-as-Judge to establish an AI quality baseline and proactively identify reliability issues before production.
  • Architected a mock MCP (Model Context Protocol) server with GitHub Actions CI/CD integration. Public portfolio code demonstrates infrastructure ownership and systems designed to prevent bugs.
  • Delivered a fully containerized performance testing suite using Docker and Locust, with enforced p95/SLA thresholds so early-stage teams can validate scalability without dedicated DevOps.
  • Partnered directly with developers to integrate testing frameworks into shipping cycles, reducing feedback loops and accelerating release velocity.

QA Test Analyst / Senior Test Engineer / Project Lead
iTech Labs (GLI Subsidiary) & Gaming Labs International, Australia
Nov 2021 – Aug 2025

  • Designed/executed 50+ structured system test plans for online gaming/wagering platforms with 100% alignment to global regulations across jurisdictions.
  • Delivered functional, regression, and compliance testing; improved defect detection by 30%.
  • Interpreted regulatory frameworks to define test scope, increasing coverage completeness by 25%.
  • Managed end‑to‑end certification projects across global teams; logged/analysed defects in JIRA.
  • Produced certification reports for clients/regulators; introduced templates that reduced manual errors.

⚙️ Skills

💻 Languages

  • Python
  • Java
  • Playwright
  • Pytest
  • Locust
  • Selenium WebDriver
  • pandas

📋 Testing & Quality

  • UI Test Automation
  • API Testing (CRUD, Schema)
  • Performance Testing (SLA)
  • AI Agent Validation
  • Data Validation (MongoDB)
  • Contract Testing (Pact)
  • Mobile Testing
  • Security Testing (OWASP ZAP)
  • Data Quality
  • Prompt Engineering
  • AI Quality Operations

🛠 Tools & CI/CD

  • GitHub Actions
  • Docker
  • Flask
  • Allure Reports
  • JSON Schema
  • OWASP ZAP
  • Railway (deployment)

📚 Relevant Coursework

UI Automation (Playwright)API Testing (REST/Schema) Performance Testing (Locust)AI Agent Testing Data Validation (MongoDB)CI/CD (GitHub Actions) Contract Testing (Pact)Selenium WebDriver (Java) Mobile Testing (Playwright)Device Emulation Security Testing (OWASP ZAP)Docker Data Quality (pandas)pytest Multi‑Agent Adversarial TestingPrompt Engineering AI Quality OperationsOrchestration & Parallel Testing
Ankit Kalluraya

MELBOURNE, AUSTRALIA

Let’s talk about your next project.

Let’s talk quality engineering, AI safety, and useful applications.

Email Ankit

GitHub · LinkedIn