Maharshi Soni

Bangalore, India | soni.maharshi1999@gmail.com | +91 7096325300 | LinkedIn | GitHub | Portfolio

Guidewire automation QA consultant with 5 years of test-automation experience designing and scaling JavaScript-based UI/API automation frameworks (TestCafe, Cucumber, Selenium) for PolicyCenter, BillingCenter, and ClaimCenter across 15+ insurance lines. Built and customized GTUI, a data-driven automation product, for a major US Guidewire client, and led a 20-person team delivering a broader automation SaaS platform that cut coding effort by 75% and lifted automation coverage by 80%. Guidewire InsuranceSuite Analyst Associate Certified, with hands-on experience integrating automated suites into CI/CD pipelines, interpreting BDD/Gherkin requirements, and mentoring 150+ engineers across global delivery teams.

Skills

Languages

PythonJavaScriptJavaSQLHTMLGOSUCC++JSON

Insurance & Domain

P&C InsurancePolicy CenterBilling CenterClaim CenterPC/BC/CCInsurance

Testing & Automation

TestCafeCucumberSeleniumTestNGQA TestingBDD/GherkinSDLCIgnite (Page Objects)

CI/CD & DevOps

CI/CDGitBitbucketTeamCityAzure DevOpsAzure ReposDockerJIRA

Product & Management

Product StrategiesProduct AnalyticsProduct RoadmapProduct PrototypeProduct PositioningStakeholder ManagementTeam ManagementAgile & Lean MethodologiesDesign ThinkingSWOT AnalysisA/B TestingMVP & MDPBacklog PrioritizationGTM StrategyConjoint & Cluster AnalysisQualitative & Quantitative Analysis

Tools & Platforms

VS CodeIntelliJGitJira & SQLTableauRESTSOAPAPI

Soft Skills

CommunicationProblem SolvingSalesProduct Positioning

AI/ML Skills (from projects)

Data Processing

NumPypandasscipypyarrow

Machine Learning

scikit-learnONNXtorch

Tools & Libraries

pydanticclickpytestJinja2pyyamlhttpxrichPlaywrightSQLite / FTS5pytest-covnetworkxbehave (BDD)Beautiful Souppdfplumberpython-docxZerodha Kite ConnectAlpaca APIpytest-playwrightbeautifulsoup4scikit-imagerequestscoloramapython-multipartwebsockets

NLP & LLM

nltksentence-transformersAnthropic ClaudeNVIDIA NIMLlama 3.1/3.2

Web & API

FastAPIFlaskuvicornStreamlitstrawberry-graphql

Visualization

matplotlibseaborn

MLOps & DevOps

Docker

Vector DB & Search

FAISSChromaDBfaiss-cpu

Computer Vision

Pillow

Employment History

Senior Guidewire Consultant - Automation QA PwC Acceleration Centers | Bangalore (08/2021 - Present)
  • Automation framework & product development: Designed and built two JavaScript, data-driven automation SaaS testing products using TestCafe and JSON-based test data, generating over $3M in revenue and sold to 15 clients; led a 20-person team through the full product roadmap
  • GTUI customization: Customized the GTUI automation testing product for a major US-based Guidewire client on PolicyCenter, earning commendation from partners, managing directors, and client stakeholders and contributing to 3 subsequent business wins
  • Framework engineering: Identified and closed 10+ critical framework gaps spanning reporting, data-driven capability, logger functionality, and synchronization, improving overall automation system performance by 40%
  • Test coverage & delivery: Authored and maintained 1,000+ test cases and 30,000+ lines of JavaScript automation code across 100+ PolicyCenter user stories, reducing production bugs by 30% while participating in all Scrum ceremonies
  • CI/CD & version control: Integrated automated regression and smoke suites into TeamCity and Azure DevOps (Azure Repos) pipelines running on containerized (Docker) build agents, using Git and Bitbucket for version control and code review, and JIRA to track quality and defect metrics
  • Page Object generation: Used Ignite to generate and refine Page Objects, then extended them with custom JavaScript functions to cover complex end-to-end and regression scenarios across web and API layers
  • Execution & stability: Executed and debugged automation runs, distinguishing application defects from script issues, and proactively fixed and refactored flaky tests to keep the suite stable over time
  • BDD/requirements translation: Used Cucumber to translate Gherkin/BDD-style requirements into maintainable, reusable automated test suites across PolicyCenter, BillingCenter, and ClaimCenter
  • Domain breadth: Automated and tested 15+ personal, commercial, and health insurance lines including Workers' Compensation, Auto, Homeowners, Farmowners, Commercial Package Liability, Businessowners, Boat, Healthcare, and Life Sciences products
  • QA operations: Drove automated smoke and regression test execution, production-support validation, and UAT support as a core member of the PolicyCenter automation QA team, identifying and tracking numerous defects to resolution
  • Mentoring & enablement: Delivered automation product training to 150+ associates through senior managers across 3 global locations, driving a 30% increase in product adoption; supported 10+ clients and delivered 15+ sales/technical demos
  • Stakeholder & delivery management: Managed 10+ stakeholders and gathered 50+ requirements using Lean methodologies and Design Thinking, reducing development time by 40%
  • Go-Live documentation: With a project manager mindset, distilled a 2-year project into a comprehensive article for partners within 2 months, selected for Go-Live
  • Process improvement: Drove 25+ business process improvements and 10+ integrations for PolicyCenter, delivering measurable cost savings and efficiency gains for clients and the firm
  • Recognition: Top 10 finalist, firm-wide Gen AI innovation challenge for generating automation test data using ChatPwC AI; Guidewire Insurance Suite Analyst Associate Certified
Subject Matter Expert - Automation Center of Excellence (Additional Responsibility) PwC Acceleration Centers | Bangalore (04/2022 - 10/2024)
  • Automation strategy leadership: Appointed Subject Matter Expert in PwC's Center of Excellence for Automation and Innovation on the strength of niche automation skills and an innovative mindset, driving firm-wide automation strategy and knowledge sharing
  • Product delivery: Used SWOT analysis, market research, and disruptive innovation to build multiple SaaS, data-driven MVPs; led a 20-person team from roadmap to sale across multiple clients using Lean and Design Thinking
  • Enablement: Trained 150+ associates through senior managers globally and delivered 15+ sales demos; Gen AI innovation challenge finalist for automation test-data generation
Business Development Executive Trainee K12 Techno Services | Bangalore (03/2021 - 06/2021)
  • Brought $1,000 worth of new digital business and set a record for employer's new education-based digital business unit (AOL)
  • Managed up to 100 calls daily, with duties including helping customers sign up, customer retention rates as academic consultant
  • Worked to ensure a positive, hassle-free customer experience as a result shown 83% customer retention rate and got recognition from sales head for quality customer service

Projects

Enhancing ACKO's Car Insurance Market Penetration in Rural India 04/2024 - 07/2024 | IIMK MDP (Management Development Programme) - Indian Institute of Management Kozhikode

Undertook a comprehensive analysis to address significant challenges faced by ACKO Technology and Services Pvt Ltd. Combined both qualitative and quantitative research methodologies - conducted a survey of 200 respondents and 15 individuals for 30 product profiles to gather data on customer experience and pricing sensitivity. Applied advanced analytical techniques such as Linear Regression, Conjoint Analysis and other product analytics tactics to uncover key factors affecting customer satisfaction, pricing sensitivity, and brand recognition in rural areas. Employed Agile, Lean, and Design Thinking methodologies to guide iteration of product features. Brought three years of experience as a Technical Consultant for US insurance companies with knowledge in insurance and digital solutions.

AI/ML Projects

insurance document analyzer advanced

Insurance policies are dense, complex documents that follow domain-specific conventions. This analyzer applies a combination of **regex pattern matching**, **NLTK-based NLP**, and **scikit-learn text

nltk . scikit-learn . numpy . pytest
    rag qa knowledge base advanced

    QA teams accumulate vast collections of test plans, runbooks, regression guides, and API documentation. Finding the right information quickly is a constant challenge. **RAG QA Knowledge Base** solves

    fastapi . uvicorn . python-multipart . jinja2
      ai bug triager advanced

      Automated bug report classification and triage powered by machine learning. Feed it a bug title and description, and it instantly predicts **severity**, **category**, **affected component**, and **sug

      scikit-learn . flask . numpy . pytest
        ai api contract validator advanced

        When pointed at a live server, the tool executes every generated test, validates each response against the spec, and produces a self-contained HTML compliance report with pass/fail indicators, timing

        requests . pyyaml . pytest . pytest-cov
          flaky test detector advanced

          Flaky tests -- tests that non-deterministically pass or fail without code changes -- silently erode confidence in CI pipelines. They waste developer time, mask real regressions, and slow down release

          numpy . scikit-learn . pytest
            bdd scenario generator advanced

            Writing BDD scenarios by hand is time-consuming and error-prone, especially when dealing with large backlogs. **BDD Scenario Generator** bridges the gap between product requirements and test automatio

            nltk . pytest
              ai visual regression tester advanced

              Traditional visual regression tools flag _that_ something changed but rarely explain _what_ changed. This project closes that gap. Given a pair of screenshots (baseline vs. current), the pipeline:

              Pillow . scikit-image . numpy . pytest
                self healing locator engine advanced

                **Automatically find alternative selectors when UI automation tests break due to changed selectors.**

                beautifulsoup4 . scikit-learn . numpy . pytest
                  langgraph qa agent advanced

                  The agent executes as a directed graph with five processing nodes and two conditional routing points:

                  sentence-transformers . torch . numpy . pytest
                    playwright ai test generator advanced

                    **Transform natural language workflow descriptions into complete, runnable Playwright test scripts.**

                    flask . pytest . playwright . pytest-playwright
                      AI Study Assistant advanced . repo

                      A RAG system enabling users to upload documents (PDF, TXT, Markdown) and receive AI-generated answers grounded in their own material with source citations. Uses semantic chunking, FAISS vector search, and Claude for generation.

                      FastAPI . FAISS . sentence-transformers . Anthropic Claude
                      • Document upload and semantic chunking
                      • Vector embedding and FAISS retrieval
                      • Question-answering with source attribution
                      RAG QA Evaluator advanced . repo

                      An automated evaluation framework for RAG chatbot systems built for P&C insurance policy documents. Evaluates answer quality across groundedness, relevance, latency, and retrieval quality using LLM-as-judge and BDD scenarios.

                      ChromaDB . NVIDIA NIM . Llama 3.1 . behave (BDD)
                      • Four-dimension quality evaluation with LLM-as-judge
                      • Local ChromaDB vector store with persistent storage
                      • BDD/Gherkin test scenarios with HTML dashboard reporting
                      Local AI QA Engineer advanced . repo

                      An autonomous AI agent that navigates websites, finds bugs, generates Playwright tests, captures screenshots, and produces structured test reports with severity ratings. Performs performance, accessibility, and responsive layout audits.

                      Playwright . NVIDIA NIM . Nemotron-120B . Llama 3.2
                      • Automated website navigation and AI-powered bug detection
                      • Playwright test code generation
                      • Multi-model support with performance and accessibility audits
                      Second Brain (AI Bookmark Manager) intermediate . repo

                      An AI-assisted bookmark manager that saves, organizes, and retrieves web pages. Fetches metadata, generates extractive summaries, creates tags, and provides semantic search with TF-IDF and cosine similarity over SQLite with FTS5.

                      Flask . Beautiful Soup . SQLite FTS5 . scikit-learn
                      • Web scraping with metadata extraction and auto-tagging
                      • Semantic search with FTS5 keyword fallback
                      • Content classification into categories
                      AI Resume Analyzer & ATS Optimizer intermediate . repo

                      A web application that analyzes resumes against job descriptions to provide ATS compatibility scoring, keyword gap analysis, and AI-powered bullet point rewrites. Supports multiple LLM providers (Anthropic, NVIDIA NIM, OpenAI).

                      Flask . pdfplumber . python-docx . Anthropic Claude
                      • ATS compatibility scoring with keyword matching
                      • AI-generated suggestions for critical keyword gaps
                      • Multi-provider LLM support for bullet point improvements

                      Education

                      Executive Product Management (eMDP) IIM Kozhikode | Remote (09/2023 - 06/2024)
                      • Overall scored 92% in this program, where I worked on 30+ real world product management applications
                      • Capstone project: Enhancing ACKO's Car Insurance Market Penetration in Rural India (see Projects section)
                      B.Tech in Information Technology SRM Institute of Science & Technology | Chennai (07/2017 - 05/2021)
                      • GPA 8/10, First Class with Distinction, in top 10% of IT batch with 33.34% Scholarship (due to top 4% in SRMJEE) for B.Tech
                      • Major project to recognize genres from tv posters with Deep learning (CNN, VCG) and got O+ with acknowledgement from HOD

                      Internships

                      Business Development Executive InCampus Pvt. Ltd. | Delhi (05/2020 - 09/2020)
                      • Led a cross-functional team of 5 Associates, found 20 Prospects and set a regional record for organization
                      • Daily identified and Nurtured the Clients via cold calling/Mail, B2B and Created brand awareness in 50 universities
                      Chatbot Developer Cognisun Inc. Pvt. Ltd. | Ahmedabad (06/2019 - 07/2019)
                      • Learned SQL, JavaScript, Google Dialog-flow, showed 10 real time deliverables within 1 month, Got recognition from CTO
                      • Developed 3 chat-bots for diverse industries like pizza delivery, hair salon, real estate, And improved revenue of firm by 10%

                      Volunteer Experience

                      Technical Consultant Pramukh Jewellers | Himatnagar (01/2013 - 02/2014)
                      • Upgraded organization with 10 disruptive digital capabilities of 30 years old legacy system, as a result revenue increased by 50%
                      • Equipped with the latest technology, firm was able to expand from 1 small town to 3 states within 3 years
                      • Improved overall efficiency in terms of payments, salary, billings, recruitment, brand image and other operational areas
                      Quality Analyst Zed Digital | Chennai (01/2020 - 02/2020)
                      • Assessed quality of application using UAT, provided approval for usability, Got appreciation from president

                      Certifications