JASWANTH

MUDAPAKA

[ Hyderabad, India · Available for hire ]

From test case to AI pipeline, I ship both. I graduated not knowing which lane to pick, so I learned all three. I break things on purpose, build systems that think, and ship code that holds

SDET

Playwright, Pytest, POM, CI/CD

Gen AI

LangChain, Bedrock, RAG, LLMs

Cloud

AWS Bedrock, S3, Lambda, EC2

TECH STACK

Languages

Python JavaScript SQL

Test Automation

Playwright Pytest Page Object Model REST API Validation GitHub Actions CI/CD

QA Practices

Functional Testing Regression Testing Exploratory Testing Root Cause Analysis Defect Documentation

AWS Cloud

Bedrock S3 Lambda EC2 CloudWatch IAM

Gen AI & LLMs

LangChain LangGraph Streamlit Prompt Engineering RAG Pipelines

AI Tools

Claude (Anthropic) GitHub Copilot Amazon Bedrock

AI / ML

Generative AI NLP Scikit-learn YOLOv5 Computer Vision

Tools

Git GitHub Docker Jupyter Notebook VS Code

Databases

MySQL

WORK EXPERIENCE

2023 – 2026 Hyderabad, India

Full-Time

QA Automation Engineer

Blue Button (Home Automation Company)

  • Built and maintained a Playwright + Pytest automation framework from scratch, reducing regression cycle time by 55–65% and enabling consistent cross-browser test coverage across the web dashboard.
  • Designed 200+ structured test cases covering lighting control, security surveillance, and AV integration workflows, establishing a reliable regression baseline across every release cycle.
  • Identified 30+ defects across functional and regression cycles, including device-trigger misfires, UI state inconsistencies, and automation response delays, reducing production escape rate by 35–45%.
  • Performed REST API testing for device-to-controller communication flows, validating request payloads, response accuracy, status codes, and failure behaviour under edge conditions such as dropped connections and invalid inputs.
  • Conducted exploratory testing on home theater and surveillance integration flows, covering real-world scenarios including network latency, concurrent device triggers, and hardware-software handshake failures.
  • Collaborated with developers and the integration team in Agile sprints to review requirements, define acceptance criteria, and ensure zero-defect releases across smart home modules.

FEATURED PROJECTS

E2E Automation

Playwright OpenCart Framework

90%+

Code coverage across 15+ workflows

Production-grade Playwright + Python automation framework with Page Object Model covering cart, checkout, auth, search & wishlist. Caught 3 browser-specific rendering defects. Zero flaky failures after resilient locator refactor.

60+

Test cases — functional, regression, BVA

PlaywrightPython PytestPOM GitHub ActionsCI/CD
View on GitHub ↗

Gen AI · LegalTech

IntakeIQ

5

Pipeline stages — Parse → Classify → Extract → Route → Alert

Intelligent document intake pipeline for law firms and claims departments. Classifies inbound docs (Insurance Claims, Court Filings, Subpoenas, etc.), extracts type-specific fields, and generates DMS-style routing paths — automating what legal teams do manually.

9+

Document types supported — zero-shot LLM classification

LangChainGroq LLaMA 3.3 70BPydantic Streamlitpdfplumber PytestGitHub Actions
View on GitHub ↗

Gen AI · QA

SmartHire — SmartHire AI

30+

Structured QA test cases authored

ReAct-pattern agentic screener built with LangGraph + Groq. Validates LLM output determinism, scoring schema compliance, and idempotent DB writes. Fixed critical scoring regression via LangGraph state guards.

15s

Candidate scoring turnaround

LangGraphLangChain PythonSQLite StreamlitGroq
View on GitHub ↗

AWS · AI Pipeline

Call Center Summarization AI

40%

Reduction in manual review workload

End-to-end audio transcription and AI summarization pipeline on AWS. Stage-by-stage functional test checkpoints across S3 integrity, transcription fidelity, AI output schema, and CloudWatch logs. Triaged 12+ defects.

78→95%

Pipeline accuracy after defect fixes

PythonLangChain AWS S3Bedrock CloudWatchNLP
View on GitHub ↗

Computer Vision

Face Mask Detection — YOLOv5

682

Training images — 3 classes

Real-time CV system classifying With_Mask, Without_Mask & Incorrect_Mask from live video. Compared YOLOv5s vs YOLOv5x — YOLOv5s faster at similar mAP. Custom-labeled dataset using LabelImg. Trained on Google Colab.

300

Optimal epochs (best mAP)

YOLOv5Python OpenCVCNN LabelImgGoogle Colab
View on GitHub ↗

Machine Learning

Crop Recommendation System

92%

Prediction accuracy

ML-based crop prediction using Random Forest. Processed 5,000+ records with Python and SQL. Feature engineering across soil type, rainfall, temperature & humidity. Deployed with Flask REST API and HTML/CSS interface.

5K+

Records processed

Random ForestPython SQLFlaskScikit-learn
View on GitHub ↗

CERTIFICATIONS

Generative AI for Software Development In Progress

DeepLearning.AI

AWS Certified Cloud Practitioner

Amazon Web Services

2024

Prompt Engineering for Developers

DeepLearning.AI

2024

Troubleshooting and Debugging Techniques

Google

2025

AI Foundations Associate

Oracle (OCI)

2025

LET'S
BUILD
SOMETHING.

Open to SDET, and QA Engineer (Manual and Automation Testing) , roles. I build automation frameworks that don't flake, AI pipelines that ship, and infrastructure that scales. If your team needs this, that's literally me.