Mini Mariya Thomas

Built a Python tool that automatically evaluates the quality of chatbot responses using the Gemini AI API — saving QA engineers hours of manual work.

• Reads a testcase file with the prompt to be asked to Chatbot and expected results
• Uses a two-step AI pipeline - first AI acts as the chatbot, then a second AI evaluates the response
• Generates a professional Word report with the details of Chatbot responses, and the evaluation results
• Report includes an evaluation summary with total test cases, pass count, fail count and quality score
• Uses the LLM-as-a-judge pattern - a real industry technique where AI evaluates AI output quality

Technologies: Python · Google Gemini API · python-docx · Prompt Engineering · python-dotenv

Built a Python tool that automatically generates structured test cases from any requirements document (BRD, User Story, or plain text) using the Google Gemini AI API - reducing manual test case writing from 3-5 hours to under 60 seconds.

• Reads any requirements document as input - BRD, User Story, or plain text
• Uses prompt engineering to instruct AI to generate positive, negative, boundary and edge test cases
• Parses AI response and exports 50+ structured test cases to a ready-to-use Excel file
• Demonstrates how AI can eliminate repetitive manual work in the testing lifecycle

Technologies: Python · Google Gemini API · OpenPyXL · Prompt Engineering

Built a Python tool that automatically generates a complete Playwright test framework from a URL and test scenarios — including page objects, test data, and test scripts using the Google Gemini AI API. Reduces framework setup from 4–6 hours to under 2 minutes.

• Accepts only a URL and test scenarios as input - no locators or manual setup needed
• Automatically fetches the page HTML and extracts interactive elements using BeautifulSoup
• AI generates real locators, a Page Object Model class, test data JSON, and Playwright test scripts
• Uses a 3-step AI pipeline - test data first, then page object, then test script — ensuring consistency across all files
• Includes a setup script that automatically installs Playwright and configures the test environment
• Generated scripts follow enterprise standards with data-driven tests, Allure annotations, and console logging

Technologies: Python · Google Gemini API · BeautifulSoup4 · Playwright · Prompt Engineering · python-dotenv

Built a Python tool that automatically analyzes the impact of code changes on existing test cases using the Google Gemini AI API — reducing manual regression impact analysis from 2-3 hours to under 60 seconds.

• Accepts a change description and existing test cases as input
• AI identifies which test cases need updating, re-execution, or removal
• Generates a professional Word report with High/Medium/Low risk areas and specific TC recommendations
• Produces an Excel file showing only the impacted test cases with Before → After changes
• AI automatically discovers edge cases not mentioned in the change description

Technologies: Python · Google Gemini API · OpenPyXL · python-docx · Prompt Engineering · python-dotenv

Built a Python tool that automatically analyzes defects and generates structured bug report from any input document using the Google Gemini AI API - reducing manual bug report writing from 20-30 minutes to under 60 seconds.

• Reads any bug document in layman language as input in a text file
• Uses prompt engineering to instruct AI to generate a structured bug report with all the relevant fields
• AI will instruct to add some missing information to make bug report more detailed
• Parses AI response and export to a ready-to-use Word document
• Demonstrates how AI can eliminate repetitive manual work in the testing lifecycle

Technologies: Python · Google Gemini API · GenAI · PythonDOCX · Prompt Engineering

Developing a feedback-directed framework to simulate adversarial input generation and uncover system vulnerabilities to support efficient and scalable penetration testing and performance testing processes.

• Automates grammar-based input generation.
• Analyzes runtime/memory metrics and updates grammars iteratively.
• Generate more inputs that serves specific run-time and memory
• Analyse the metrics and user defined properties for various functions and versions to analyze the best performed functions/versions

Technologies: Python · Fandango · JSON · WeasyPrint

• Designed and implemented end-to-end UI test automation for the SauceDemo web application (https://www.saucedemo.com) using Playwright and JavaScript.
• Implemented environment-based execution (QA/Staging/Prod) with dynamic configuration handling.
• Applied the Page Object Model (POM) and structured logging to create reusable and maintainable test components, covering login, inventory, cart, checkout, filters, and navigation flows.
• Integrated automated test execution into a Jenkins CI pipeline with headless execution, generating both Playwright HTML reports and Allure interactive reports for enhanced test visibility.
• Configured Allure reporting to provide severity tagging, feature categorization, failure analysis, and screenshot attachments for enterprise-level reporting.
• Enabled fully parallel test execution with CI-optimized worker strategy.
• Version-controlled the framework using Git and GitHub.

Technologies: Playwright · JavaScript · Jenkins · CI/CD · GitHub · Allure Reporting · HTML Reports · Page Object Model · Environment Config

• Automated multiple pages of the application (https://groceryapp.uniqassosiates.com) using Java and Selenium
• Pushed the latest updates to GitHub for version control
• Integrated automated test execution into CI/CD pipelines using Jenkins.

Technologies: · Java · Selenium · Jenkins · Github · Version Control · Test Automation · TestNg · Maven

• Given a list of port numbers, we want to use this list to see if the local port is listening.

• Given an email folder contains copies of the emails John, Martha, and Mary have been sending to each other(and themselves). Find the file containing a password.

• Determine IP ranges to scan (in this case it will be localhost) and you can provide the input in any way you want
• Determine the port ranges to scan
• Determine the type of scan to run (in this case it will be a simple TCP Connect Scan)

• Configured attacker (Kali) and target (Ubuntu) VMs using VirtualBox.
• Enabled remote access and tested network communication.

Technologies: VirtualBox · Kali Linux · Ubuntu · Scanning

• Features: List, Add, Edit, Delete tasks
• Concepts: Lists, loops, functions, file I/O
• Save tasks to a text file
Click for Script

• Created a bash script that parses /var/log/auth.log to count failed/successful logins.
• Extracted IPs from failed attempts and produced summary reports.

Technologies: Linux · Bash

• Task 1: Enumerate Subdomains- Add subdomains to a list, Make it a http url, and Send request and check the subdomain exists. Click for Script
• Task 2: Build a simple directory enumeration tool using Python. Click for Script
• Task 3: Build a simple ICMP scanner to identify potential targets on the network using Python. Click for Script | Click to see Output
• Task 4: Build a simple port scanner to identify open scripts using Python. Click for Script

Technologies:· Python · Powershell · Shell script · Programming

• Solved 10+ SQL injection challenges and retrieved hidden data.
• Performed blind SQLi and UNION attacks with Burp Suite.

Technologies: SQL Injection · Burpsuite · Documentation

• Completed 10+ PortSwigger labs covering reflected, stored, and DOM XSS.
• Exploited HTML/JS contexts, bypassed encoding, and used Burp for testing.

Technologies: XSS · Burpsuite · JavaScript

• Monitored security alerts using SIEM tools.
• Investigated firewall and email logs with Splunk, Elastic, and Wireshark.

Technologies: Log Analysis · Splunk · SIEM · Wireshark

Simulated DDoS attack response: analyzed logs, mitigated threats, and created postmortem reports.

Technologies: Log Analysis · Firewall Rule · Communication

Performed phishing analysis and awareness design as part of Mastercard’s Virtual Experience Program.

Technologies: Phishing Simulation · Threat Analysis · Communication

Investigated suspected security breach, analyzed web activity logs, and identified automated attack patterns.

Technologies: Log Analysis · Pattern Recognition · Cybersecurity Investigation

Formally modeled IKEv2 protocol using Tamarin Prover to verify key secrecy and message agreement.

Technologies: Tamarin Prover · Cryptographic Protocols

Developed grammar-based fuzzer for SQLite to test SQL coverage using Docker and coverage reports.

Technologies: Python · SQLite · Docker · Coverage Analysis

Built a system to learn input constraints from positive and negative sample sets using formal grammars.

Technologies: Python · AST Parsing · Blackbox Oracle Testing