Results-driven Software Developer & SDET II with 3+ years building enterprise-grade
automation frameworks, full-stack distributed platforms, and AI-assisted tooling at GreyOrange.
Architected a data-driven Playwright suite covering 147 E2E scenarios and contributed to a
warehouse-simulation platform running 30 concurrent Celery workers with GCP analytics.
Solved 1,000+ LeetCode problems. Passionate about applying AI/ML & agentic tools
to build faster, smarter, and at scale.
Professional Experience
Software Development Engineer in Test II (SDET II)
GreyOrange Pvt. Ltd. · Warehouse Robotics & Fulfillment Automation
- Architected GSB Automation Suite — a zero-mock, data-driven Playwright (TypeScript) E2E framework for a PixiJS canvas map editor; grew to 147 JSON-driven test scenarios across Barcode, PPS & Handheld entities, asserting Redux store state instead of visual pixels.
- Designed an entity-agnostic registry architecture (action + entity registries, topological dependency sorting) so new entity suites need only JSON — zero engine code changes — slashing new-suite setup time by ~70%.
- Built 33 reusable Playwright helpers (canvas world→screen coordinate mapping, Redux state readers, carousel assertions) adopted across the entire QA team; 38 registered action handlers.
- Integrated Google Cloud Build CI/CD pipeline with automated JUnit XML + Playwright HTML reporting and email notifications — enabling fully unattended nightly regression runs.
- Developed a Jira integration module within the GSB Simulator (Django REST Framework) to auto-create and link Jira tickets from simulation failure events, reducing manual triage overhead.
- Contributed to the full-stack GSB Simulator platform (Django + Celery + RabbitMQ + PostgreSQL + Next.js + Three.js) processing parallel simulations via 30 concurrent workers, with GCP BigQuery analytics, real-time WebSocket updates, and Keycloak auth.
- Leveraged AI-assisted development (Cursor, Claude Code) for accelerated feature delivery, embodying modern agentic engineering — from prompt-driven scaffolding to AI-reviewed PRs.
- Applied pytest + Selenium automation for earlier validation layers and legacy browser compatibility across the product suite.
Key Projects
GSB Automation Suite — Canvas E2E Testing Framework
Playwright · TypeScript · Jenkins · Redux
Entity-agnostic, data-driven test engine for a PixiJS warehouse map editor. 147 scenarios defined in JSON; engine auto-discovers & executes with topological ordering. Assertions via Redux store injection (no pixel matching). Includes 33 shared helpers, 38 action handlers, and full CI email reporting. Architecture proven by adding PPS suite (111 cases) with zero engine changes.
GSB Simulator — Warehouse Simulation & Optimization Platform
Django · Celery · RabbitMQ · Next.js · Three.js · GCP · scikit-learn
Full-stack platform for physics-based warehouse automation simulations before real deployment. 30-worker Celery + RabbitMQ distributed backend, PostgreSQL, WebSockets for live updates, Three.js 3D visualization, LightGBM/scikit-learn ML models, BigQuery analytics, Keycloak auth, and custom Jira ticket automation. Kubernetes-ready.
Achievements
1,000+
LeetCode Solved
DSA: arrays, graphs, DP, trees, backtracking
147
Automated Scenarios
JSON-driven E2E, zero pixel mocking
~70%
Setup Reduction
New suites: JSON only, no engine code
Personal Projects
Logcat Intelligence Engine — Self-Improving AI Diagnostic System
Python · Qwen2.5-7B · QLoRA · FAISS RAG · Docker · vLLM
Self-improving diagnostic system for Android OS logs. A ReAct agent investigates logcat/dmesg/bugreport artifacts using RAG-backed tools; traces feed SFT+DPO data to fine-tune a local Qwen2.5-7B model via QLoRA, gated by an eval harness and served through vLLM in Docker. Baseline: 90% diagnostic accuracy.
Baklola E-Commerce — AI-Vibe-Coded Full-Stack Application
React · Cursor · Claude AI · Vibe Coding
End-to-end e-commerce app built entirely via AI-assisted (vibe coding) workflow using Cursor & Claude — demonstrating rapid prototyping ability, modern agentic tooling proficiency, and the skill to go from idea → working product with AI as a co-pilot.
Interests & Continuous Learning
AI/ML & Agentic Systems — Actively building with LLMs, RAG pipelines, Claude API & agentic workflows |
Competitive Programming — 1,000+ LeetCode; strong algorithmic thinking |
System Design — HLD/LLD, distributed systems, scalable architecture |
Open Source & Shipping — Always learning, always building on GitHub