📚 Currently Learning

🐍Python for AI◆ 🤖GenAI & LLMs◆ 🗂️RAG Pipelines◆ 🔗LangChain◆ 💻GitHub Copilot◆ ⚡AutoGen◆ 🔌MCP Protocol◆ 🎯Vibe Coding◆ ✍️Prompt Engineering◆ 🔍Vector Databases◆ 🧪AI Agent Testing◆ 📊LLM Evaluation◆ 🛡️Hallucination Detection◆ 📈RAGAS Framework◆ 🐍Python for AI◆ 🤖GenAI & LLMs◆ 🗂️RAG Pipelines◆ 🔗LangChain◆ 💻GitHub Copilot◆ ⚡AutoGen◆ 🔌MCP Protocol◆ 🎯Vibe Coding◆ ✍️Prompt Engineering◆ 🔍Vector Databases◆ 🧪AI Agent Testing◆ 📊LLM Evaluation◆ 🛡️Hallucination Detection◆ 📈RAGAS Framework◆

page.goto("https://example.com")

@pytest.fixture(scope="session")

from playwright.sync_api import Page

expect(page).to_have_title()

def test_login(page: Page):

page.locator("#username").fill()

import pytest

page.click("button[type='submit']")

assert response.status == 200

page.wait_for_selector()

from pages.login import LoginPage

expect(element).to_be_visible()

Open to SDET / AI Tester Roles

Soumyaranjan Sahu

Software Development Engineer in Test · AI-First QA Engineer

4.6+ years building quality into software — from precision manual testing to AI-powered automation frameworks that reduce test creation effort by 70%.

Playwright Python LLM Integration RAG Testing Agentic AI Prompt Engineering CI/CD ISTQB

View My Work → ⬇ Resume 💼 LinkedIn 🐙 GitHub

0Years in QA

0Effort Reduced

ISTQBCertified

🐍 ai_test_generator.py

# 🤖 AI generates test from plain English

from llm_agent import TestGenerator

from pages.login import LoginPage

scenario = "user logs in → dashboard"

test = TestGenerator.generate(scenario)

def test_valid_login(page):

expect(page).to_have_url("/dashboard")

✔ PASSED · 1.24s · Zero manual coding

01 — Technical Stack

Skills & Expertise

Test Automation

PlaywrightPythonPytestPage Object Modelfixturesconftest.pyassertionslocatorscross-browser

Manual QA

Functional TestingRegressionUATIntegration TestingSystem TestingSmoke & SanityBRD/FRD ReviewRTM

Manual Testing95%

Test Case Design92%

STLC / SDLC90%

API Testing

PostmanPlaywright APIRequestContextREST APIStatus CodesPayload ValidationJSON SchemaSQL

Postman88%

REST API Testing85%

SQL Validation72%

AI / LLM Testing

LLM IntegrationAgentic AIRAG TestingPrompt EngineeringAI Test GenerationAllure ReportsPytest-HTML

Agentic AI Workflow75%

Prompt Engineering70%

Allure Reports82%

CI/CD & Tools

GitHub ActionsGitJIRAConfluenceYAML WorkflowsAgile / Scrum

GitHub Actions80%

Git78%

JIRA92%

AI Testing Roadmap 🚀

Skills I am actively learning to become a future-ready AI Tester — from GenAI fundamentals to agentic frameworks.

🐍 Python for AI Testing ✓ Strong

Core Python, OOP, scripting for test automation and AI integrations. Used daily in Playwright frameworks.

Python 3OOPScriptingPytest

🤖 GenAI & LLMs Learning

Understanding Generative AI and Large Language Models — how they work, how to test them, prompt engineering, and output validation.

GenAILLMsGPT-4ClaudeGeminiPrompt Engineering

🗂️ RAG (Retrieval-Augmented Generation) Learning

Testing RAG pipelines — retrieval accuracy, context relevance, vector DB validation, chunking strategies, and hallucination detection.

RAGVector DBEmbeddingsChromaDBPinecone

💻 GitHub Copilot Learning

Using AI-powered coding assistant to write test scripts faster, generate boilerplate POM code, and auto-complete Playwright test cases.

GitHub CopilotAI Pair ProgrammingCode Generation

🔗 LangChain Next Up

Building and testing LLM-powered applications using LangChain — chains, agents, tools, memory, and evaluations for QA pipelines.

LangChainChainsToolsMemoryLLMOps

⚡ AI Agents & AutoGen Next Up

Testing multi-agent systems built with AutoGen — validating agent collaboration, decision-making, tool usage, and error recovery in autonomous workflows.

AutoGenAI AgentsMulti-AgentTool UseAutonomous

🔌 MCP (Model Context Protocol) Next Up

Testing MCP servers and clients — validating tool integrations, context passing, and AI-to-tool communication in agentic applications.

MCPModel Context ProtocolTool IntegrationAnthropic

🎯 Vibe Coding (Vibium) Exploring

AI-assisted development and testing using natural language-driven coding tools — write tests by describing behaviour in plain English.

VibiumVibe CodingNL-to-CodeAI-first Dev

02 — Future-Ready

AI Tester Capabilities

The next generation of QA is at the intersection of AI and testing. From validating LLM outputs to testing RAG pipelines — here's where the journey is heading.

✓ Production🤖

Agentic AI Test Generation

Production LLM-powered framework converting plain English to Playwright Python scripts. Zero manual coding — 70% effort reduction at Cozentus / DP-World Connect.

LLM AgentPlaywrightPytestPOMPrompt Templates

Learning🔍

LLM Output Validation

Testing LLM outputs — checking hallucinations, response consistency, factual accuracy, and expected formats across prompts.

LLMHallucination DetectionOutput Validation

Learning🗂️

RAG Pipeline Testing

Validating Retrieval-Augmented Generation systems — retrieval accuracy, context relevance, vector database queries, end-to-end pipeline correctness.

RAGVector DBEmbeddingsRetrieval Testing

✓ Active✍️

Prompt Engineering for QA

Crafting precise prompt templates instructing LLMs to generate structured, production-ready test code following coding standards.

Prompt EngineeringFew-shotChain-of-thought

Next⚡

AI Agent Testing

Testing multi-step AI agents using tools, APIs, and memory. Validating decision-making, tool-use accuracy, error recovery across complex workflows.

AgentsTool UseLangChain

Next📊

AI Model Evaluation

Systematic evaluation using benchmarks, evals frameworks, and custom test suites. Measuring accuracy, latency, cost, and regression testing model upgrades.

EvalsRAGASBenchmarks

03 — Work History

Experience Timeline

⏱ Years of Experience

4.6+ Years

Meditab (3 yrs) + Cozentus (1.4 yrs) + More

May 2021 → Present

Current

Oct 2024 — Present

Senior Quality Engineer

Cozentus Technologies Pvt Ltd · Bhubaneswar

AI-powered framework — LLM converts plain English → Playwright scripts, 70% effort cut
Automated test cases using Pytest + POM design pattern
Maintained POM classes (LoginPage, CartPage, CheckoutPage) across UI changes
REST API testing via Postman — payloads, status codes, business rules
SQL queries for backend data validation & database verification
Analyzed Allure HTML reports to identify failures and root causes
Agile Scrum — Sprint Planning, Stand-ups, Reviews, Retrospectives

Oct 2021 — Sep 2024

Quality Engineer

Meditab Software Pvt Ltd · Ahmedabad

Reviewed BRD, FRD, and user stories to understand workflows and business logic
Designed test cases — positive, negative, boundary, edge scenarios
Functional, Regression, Integration, and E2E testing across multiple releases
UI and cross-browser testing for consistent user experience
API validation using Postman — request/response payloads and business rules
Maintained RTM and test execution reports; release smoke testing
Defect triage with developers and BAs for root cause analysis

May 2024 — Nov 2024

Software Testing Mentor

StarAgile · Remote

Mentored aspiring QA engineers on manual & automation testing concepts
Conducted sessions on Selenium, test case design, and defect lifecycle
Guided students through real-world testing scenarios and best practices

May 2021 — Sep 2021

Intern

RailYatri · Remote

Gained hands-on experience in software testing methodologies
Assisted in functional and regression testing of web applications
Learned industry-standard QA processes and defect tracking workflows

May 2020 — Present

Digital Creator

YouTube · Globo Odisha · 44K+ Subscribers

Created and managed a YouTube channel with 44,000+ subscribers
Produced engaging digital content — video editing, scripting, and publishing
Built community engagement and grew audience through consistent content strategy

04 — AI Work

🤖 Agentic AI Framework (Python + Playwright)

Production AI-powered test automation system where natural language test scenarios are processed by an LLM agent to automatically generate production-ready Playwright Python scripts.

Write Plain English Scenario

"User should login with valid credentials and be redirected to dashboard"

LLM Agent Processes Input

Prompt template engine generates code following POM structure, Playwright locators, and Pytest fixtures

Production-Ready Script Generated

Complete Pytest test function with POM class, locators, fixtures — zero manual coding

Execute & Analyze with Allure

Test suite runs via Pytest, analyzed in Allure HTML reports with failure root cause identification

🐍 PYTHON + 🎭 PLAYWRIGHT // AUTO-GENERATED BY LLM AGENT · ZERO MANUAL CODING

# Input: "user logs in with valid creds"
from pages.login_page import LoginPage
from playwright.sync_api import expect

def test_valid_login(page, base_url):
  login = LoginPage(page)
  login.navigate(base_url)
  login.enter_username("admin")
  login.enter_password("admin123")
  login.click_submit()
  expect(page).to_have_url(f"{base_url}/dashboard")

✔ PASSED · 1.24s · AI-Generated

70%

Reduction in test creation effort — measured in production at Cozentus Technologies on DP-World Connect Application

05 — Portfolio

Projects & Builds

AI · Featured

AI-Powered Test Framework — DP World Connect

Production framework where natural language scenarios are auto-converted to Playwright Python scripts via Agentic AI (LLM). Maintained POM classes, analyzed Allure reports, reduced test authoring effort by 70%.

PythonPlaywrightLLM/Agentic AIPOMPytestAllureJIRA

Internal — Cozentus →

Automation

E-Commerce UI Automation Suite

End-to-end Playwright + Python test suite with full POM architecture covering login, cart, checkout, and payment flows.

PlaywrightPytestPOMfixtures

GitHub →

API Testing

REST API Automation Suite

API test suite using Playwright's APIRequestContext — validates CRUD, auth flows, response schemas, and error handling.

APIRequestContextJSON SchemaPytest

GitHub →

CI/CD

GitHub Actions CI Pipeline

Automated CI pipeline triggering Playwright tests on every push/PR. Generates Allure HTML reports and publishes artifacts.

GitHub ActionsYAMLAllurePytest-HTML

GitHub →

Manual QA

Show IT Max — Meditab

3-year QA ownership of Meditab's ShowITMax healthcare app. Full regression, cross-browser, API validation across 10+ release cycles.

Manual TestingPostmanJIRARTM

Internal Project →

06 — Credentials

Certifications & Education

🏅

ISTQB Foundation Level

GASQ · Global Assoc. for Software Quality

2024 · Certified

▸ View Certificate

🎭

Playwright with Python

Udemy · Test Automation

2025 · Completed

🐍

Python Programming

Udemy · Core Python & OOP

2025 · Completed

🔍

SelectorsHub Certification

SelectorsHub · XPath & CSS Selectors

Completed

🧪

Software Testing (Manual + Selenium)

Certification · Manual & Selenium Automation

Completed

📂

GIT Basics (Zero to Hero)

Certification · Version Control & Git

Completed

📋

Masterclass Software Testing with Jira & Agile

Certification · Jira, Agile & Testing

Completed

🎓

B.Sc. Computer Science

Fakir Mohan College · 72.2%

2020 · Graduated

🎓

MCA — Computer Science

Sikkim Manipal University (SMU)

2025–2027 · Pursuing

Soumyaranjan Sahu

Skills & Expertise

Test Automation

Manual QA

API Testing

AI / LLM Testing

CI/CD & Tools

AI Testing Roadmap 🚀

AI Tester Capabilities

Experience Timeline

Senior Quality Engineer

Quality Engineer

Software Testing Mentor

Intern

Digital Creator

🤖 Agentic AI Framework (Python + Playwright)

Write Plain English Scenario

LLM Agent Processes Input

Production-Ready Script Generated

Execute & Analyze with Allure

Projects & Builds

AI-Powered Test Framework — DP World Connect

E-Commerce UI Automation Suite

REST API Automation Suite

GitHub Actions CI Pipeline

Show IT Max — Meditab

Certifications & Education

Certified TesterFoundation Level (CTFL)

OPEN TO WORK

Certified Tester
Foundation Level (CTFL)