Om
Shinde
Results-driven SDET building automated testing frameworks with Java, Selenium, and AI tools. Passionate about eliminating manual effort through intelligent automation — from OTP handling to AI-powered code generation.
About
Building reliable automation with intelligent depth
I'm an Automation Engineer at QualityKiosk Technologies, Mumbai, working at the intersection of test automation, AI tooling, and CI/CD engineering. My work is mostly about one thing: taking the manual, repetitive, error-prone parts of quality assurance and turning them into systems that just run.
I graduated with a B.Tech in Computer Science (CGPA 8.9) from D. Y. Patil College of Engineering & Technology, Kolhapur in 2024. Since then I've built API frameworks from scratch, shipped a fully autonomous OTP-handling pipeline, and grown increasingly obsessed with how LLMs and RAG can make testers dramatically more productive without sacrificing correctness.
I like hard automation problems, clean abstractions, and the quiet satisfaction of a green pipeline that a whole team trusts.
What I care about
- Scalable, reusable automation frameworks
- AI-driven test generation & code quality
- CI/CD integration and pipeline reliability
- Eliminating manual effort through smart tooling
- RAG systems for developer productivity
Education
Experience
Roles where I built impact at scale
How I Work
Five principles behind every framework I build
Understand the system first
Before writing a single test, I map the real user flows, the auth boundaries, and the parts of the app that break most. Good automation starts as good product understanding — not as code.
Design for reuse, not for speed
I build frameworks around the Page Object Model, shared utilities, and data-driven strategies. The goal is that the tenth test costs a fraction of the first — and that a junior can extend it without breaking things.
Kill flakiness at the root
Flaky tests destroy trust in automation. I isolate environment state, add deterministic waits and retries, and treat every intermittent failure as a real defect in the harness until proven otherwise.
Wire it into the pipeline
Automation that runs on a laptop is a demo. I integrate suites into Jenkins / GitHub Actions so they run on every release, surface clear reports, and give the team real confidence to ship faster.
Let AI remove the boilerplate
I use LLMs and RAG deliberately — grounded in the actual framework source — to generate POM-aligned test code, catch edge cases, and hand testers superpowers without hallucinated methods.
Projects
Flagship work across automation and AI
QA Forge: The Digital Hammer of Quality
Enterprise-grade AI manual test case generator for Jira. Automates the creation of test cases from Jira ticket descriptions using Groq LLMs, with deep Jira integration for subtask creation and Excel exports.
- Built a 'Neural Analysis Engine' using Groq's Llama-3 to analyze Jira descriptions and identify hidden edge cases.
- Implemented 'Ironclad Security' using AWS Cognito Hosted UI and RSA-signature verification for all API calls.
- Designed 'Deep Jira Integration' for smart fetching, auto-subtask creation, and Excel injection directly into tickets.
- Crafted a premium 'Digital Forge' UI with glassmorphism, Framer Motion animations, and universal responsiveness.
DataSync: Enterprise Policy Reconciliation Engine
Mission-critical policy reconciliation platform for enterprise insurance systems. Performs field-level data validation and reconciliation between legacy and modern policy databases, with real-time comparison analytics and automated reporting.
- Architected a 'JsonPath-Powered Mapping Engine' supporting 1,549+ field instances across POLICY, MEMBER, UNDERWRITING, and RIDER business domains.
- Built a 'Multi-Type Comparison Engine' with 5 intelligent comparison modes (EXACT, CASE_INSENSITIVE, NUMERIC, DATE, BOOLEAN) for heterogeneous data validation.
- Engineered 'HAL-Link Resolver' with parallel HTTP calls, exponential backoff retry logic, and connection pooling for high-throughput API interactions.
- Crafted 'Real-Time Dashboard' using React + Vite with drill-down analytics, field-level discrepancy tracking, WebSocket updates, and Excel export for audit compliance.
- Implemented 'Distributed OAuth2 Client-Credentials Flow' for secure inter-service authentication and API gateway integration.
- Containerized pipeline using Docker + Jenkins with automated test suites (RestAssured, Selenium) and CI/CD deployment gates.
AI-Powered Test Recorder & Code Generator
Building an intelligent system that records manual browser interactions and auto-generates production-ready Java Selenium TestNG test cases using AI, eliminating hand-written boilerplate entirely.
- Architected a 4-phase pipeline: framework setup → browser recording → AI action-to-method mapping → TestNG class generation.
- Uses ChromaDB + LLaMA 3.1 to ensure POM-aligned, hallucination-free code output.
- Bridges manual exploratory testing and automated code generation for non-technical testers.
RAG-Based Framework Intelligence Assistant
Built a Retrieval-Augmented Generation (RAG) system that makes LLMs answer questions strictly based on your automation framework's actual source code — eliminating hallucinated methods and incorrect POM assumptions.
- Implemented method-level chunking of Java framework files with semantic embeddings using SentenceTransformers (all-MiniLM-L6-v2).
- Integrated Groq's LLaMA 3.1 with strict grounding constraints — responds 'Not implemented' instead of fabricating answers.
- Designed clean RAG pipeline: source code → chunking → embeddings → Top-K retrieval → grounded LLM response.
Zero-Touch OTP Automation System
Built a fully autonomous OTP handling system eliminating manual intervention in authentication-heavy workflows. Designed a real-time pipeline integrating Android SMS capture, Firebase sync, and Selenium execution to enable uninterrupted test flows across banking and insurance applications.
- Engineered a 5-stage pipeline: UI trigger → SMS capture → Firebase push → Web retrieval → auto-injection
- Eliminated 100% manual OTP dependency in automation suites
- Handled real-time synchronization between mobile and web layers
- Designed retry + fallback mechanisms for delayed or missing OTPs
- Enabled stable execution for authentication-heavy enterprise flows
- Reduced execution breaks caused by OTP delays and flakiness
Geofencing-Based Hospitality Experience System
Developed a geofencing-based system to deliver personalized, location-aware services within a hotel ecosystem. Enabled automated guest engagement using GPS-triggered events to enhance user experience and operational efficiency.
- Designed GPS-based geofence triggers for automated check-in/check-out reminders
- Delivered real-time promotional alerts and nearby facility recommendations
- Built a role-based admin panel to manage geofence zones and user activity
- Implemented live tracking and event logging for monitoring guest movement
- Improved customer engagement through context-aware notifications
Skills
Tools, systems, and capabilities
Languages
Test Automation
Tools & DevOps
AI & GenAI
Databases
Writing
Notes on testing, frameworks & AI
The architecture decisions behind a scalable RestAssured + TestNG framework.
Why retrieval beats fine-tuning for hallucination-free test-code generation.
Practical patterns for deterministic, trustworthy test suites in CI.
Awards
Signals of excellence and ownership
Recognized for exceptional delivery on the ICICI Lombard General Insurance project — specifically for building a scalable API Automation Framework from scratch that directly improved client confidence, regression reliability, and release velocity.
Awarded as a standout new joiner for exceptional ramp-up speed, proactive contributions to test automation, and demonstrating ownership of real-time project deliverables within the first months of joining.
Contact
Open to new opportunities
Reach out for roles in automation engineering, AI testing, or simply a conversation about intelligent test frameworks and developer tooling.