Akarsh Mahamuni

SDET | AI Testing Engineer with 4 years of experience in manual and automation testing using Selenium, Playwright, Python, and API testing.

Specialized in GenAI and RAG systems, including LLM evaluation, AI-driven test generation, and end-to-end QA automation workflows

SDET & GenAI Automation Engineer — AI-Powered End-to-End QA Automation Platform at Celsior Technology (2025-04 – Present)

• Architected a full-stack AI-driven QA system automating the entire pipeline from Requirement Analysis → Test Case Generation → Automation Script Creation → Autonomous Execution.
• Built a multi-format RAG pipeline ingesting PDFs, Jira stories, Excel, Word, Markdown, audio, and video inputs — enabling contextual requirement intelligence and automated test generation; reduced manual test authoring time by 60%.
• Designed LLM-based structured prompting engine to transform requirements into high-quality manual test cases, improving test coverage consistency across releases.
• Developed LLM-powered script generation engine producing executable Selenium (Java/Python) and Playwright automation scripts directly from requirement-derived test cases.
• Engineered agentic AI browser automation framework integrating Browser-based AI agents, a custom Selenium autonomous framework, and Playwright MCP-based execution agent.
• Built modular FastAPI backend and Streamlit UI enabling scalable multi-agent orchestration across test lifecycle stages.

AI Quality Engineer — AI Model Solutions Testing at Celsior Technology (2024-12 – 2025-03)

• AI-driven Object Repository Model: Tested an AI-powered object repository system enabling automated web element capture via browser extension (single-element and full-page inspection) with DOM metadata and visual snapshots. Validated object detection accuracy, storage integrity, and end-to-end model training pipeline for dynamic element identification using logical names in Selenium. Verified self-healing capabilities by ensuring automatic recovery of failed locators (XPath/CSS) using AI-based object recognition.
• AI-based Page Comparison Model: Designed and executed visual regression testing using pre- and post-change full-page screenshots to detect UI differences. Validated AI model outputs for accurate change highlighting and conducted manual evaluation of false positives/negatives to improve model reliability. Tested robustness against dynamic UI changes, layout shifts, and cross-page variations.
• AI-based SRS Generation & Requirement Intelligence Engine: Designed and validated an AI-powered SRS generation system that extracts requirements from multi-modal inputs (audio, JSON, UI, tables) using LLMs and semantic similarity models. Implemented confidence scoring using cosine similarity and contextual relevance, along with requirement deduplication and traceability matrix generation. Tested end-to-end requirement extraction accuracy, consistency, and robustness across diverse data sources and edge cases.

Quality Engineer — Automation Framework Management Tool (AFMT) at Celsior Technology (2022-07 – 2024-12)

• Performed manual testing including test case design for new features, defect logging, retesting, and regression testing as part of defect lifecycle management.
• Validated ETL pipelines including File, Schema, Data, and Row Count Validation — ensuring high data accuracy across releases and cutting data defect escape rate by 35%.
• Designed scalable Selenium automation framework using POM architecture supporting parallel execution, reusable utilities, and regression suite optimization; reduced regression cycle time by 40%.
• Authored comprehensive functional and data-driven test cases through in-depth requirement analysis, driving full automation pipeline readiness.
• Managed defect lifecycle and test tracking using Azure DevOps TFS, maintaining clear traceability from requirements to test outcomes.

Bachelor of Engineering – Visvesvaraya Technological University, Belagavi