Karuna Vasu

Karuna Vasu

Software Engineer

Backend/Platform Engineer | Python | Automation Systems | AWS | AI Systems

Automation-first platforms, BotVeta orchestration, Bedrock document intelligence, and voice workflows for US Healthcare RCM.

US Healthcare RCMAzure DevOps Automation (156+ repos)BotVeta OrchestrationBedrock & Amazon Connect

About

What I build and how I think

Backend-focused Python developer in US Healthcare (RCM). I ship scalable automation, BotVeta-style orchestration, and AI workflows (Bedrock document extraction, Amazon Connect IVR) that cut turnaround time and manual effort.

What I'm best at

Automation & OrchestrationFastAPI + Next.js systemsAI-assisted IVR workflows

Current focus

Extending BotVeta — multi-tenant scheduling, monitoring, and scaling of Python and Node automation with admin and operator experiences.

Work Experience

My professional journey and key achievements

Backend / Platform Engineer (Python Developer)

Integrity Healthcare Solutions Pvt Ltd

Ahmedabad, Gujarat, India • December 2023 - Present

Healthcare RCM & automation

  • Architected scalable automation for eligibility, authorization, and claims; delivered across 156+ Azure DevOps repositories in the Automation project (payer/client workflows).
  • Built high-performance pipelines integrated with Availity, accelerating claim processing turnaround by up to ~95%.
  • Distributed bot orchestration with queues, retries, centralized logging, and fault-tolerant execution; FastAPI microservices improved reliability on key workflows by ~40%.
  • Automated EHR flows (Practice Fusion) and reusable validation frameworks, improving efficiency by ~40% and cutting regression effort by ~80%.

BotVeta (orchestration platform)

  • Bot-manager services (FastAPI + Next.js + TypeScript) with JWT RBAC, APScheduler timed runs, and full run/job lifecycle tracking across Python and Node bot servers.
  • Orchestration API on .NET 9: SQL Server, EF Core, Hangfire, Serilog, Swagger; operator UI (React 19 + Vite) and admin console (Next.js 16).
  • Enterprise integrations: Azure Key Vault, AWS credential flows, SFTP, Gmail, SQL/ODBC pipelines.

AI document & voice

  • AI PDF extraction in admin: large uploads, split/bundle workflows, FastAPI + AWS Bedrock (Claude 3.5), rule-based and confidence-aware validation, JSON/Excel outputs for RCM.
  • IVR platform on Amazon Connect with Bedrock conversational flows; Python/AWS backends (EC2, DynamoDB, Lambda) and React dashboards; Excel-driven triggers — ~60–80% less manual calling.

Research Intern

Manipal Institute of Technology

Udupi, Karnataka, India • February 2023 - December 2023

Conducted research on Electrocardiogram Heart Rate Variability Signal Analysis, contributing to published IEEE paper and earning Best Paper and Best Presenter awards.

Data Science Intern

Zephyr Technologies and Solutions PVT LTD

Mangaluru, Karnataka, India • June 2023 - July 2023

Developed a real-time chat application using Python and Django with WebSocket connections for real-time messaging and PostgreSQL for database management.

Featured Projects

Production-grade systems and platforms I've built

Bot OrchestrationUS Healthcare RCMFastAPI + Next.jsBedrock & Amazon Connect
Hero Project

BotVeta — Automation Orchestration Platform

Multi-tenant platform to schedule, monitor, and manage distributed automation workflows across organizations.

  • Orchestration APIs → scheduler → bot runners → centralized logging & monitoring
  • Queue-based execution, retries, failure recovery, JWT RBAC, mixed Python/Node backends
  • Operator UI (React + Vite) and admin (Next.js); .NET 9 API with Hangfire and SQL Server

Architecture (High Level)

Next.js admin / React-Vite operator → FastAPI & .NET 9 APIs → Hangfire / APScheduler → runners → SQL

My Role: Backend, integration, and platform features across APIs and UIs

Impact: Scalable workflow execution; strong reduction in manual coordination

Stack:PythonFastAPI.NET 9ReactNext.jsHangfireSQL ServerAWS
📄

AI-Powered PDF Extraction & Document Intelligence

Healthcare RCM document pipeline: split, group, validate, and extract structured data from large PDFs using Bedrock.

  • Admin-side upload (e.g. up to ~200MB), patient- and document-type-aware grouping
  • FastAPI (uvicorn) + AWS Bedrock; analyze-only mode; confidence-aware validation
  • JSON and Excel outputs for EOB, claims, and patient-centric workflows

Architecture (High Level)

Next.js admin → FastAPI → AWS Bedrock (Claude) → validation → exports

Impact: Faster document turnaround with auditable, rule-based + AI validation

Stack:PythonFastAPIAWS BedrockLLMsNext.js
🤖

AI Voice Automation (Amazon Connect + Bedrock)

Production IVR automation for payer calls (claims, eligibility, authorization) with AI navigation and agent handoff.

  • Amazon Connect integrated with Bedrock (Claude 3.5 Sonnet) for conversational IVR
  • Python backends on EC2, DynamoDB, Lambda; React dashboards for monitoring and control
  • Excel-driven triggers for ops; IAM-aligned secure cloud deployment

Impact: ~60–80% less manual calling; faster response turnaround

Stack:PythonAmazon ConnectAWS BedrockLambdaReact
🏥

RCM Automation Suite (Eligibility • Auth • Claims)

Rule-based automation engines for eligibility, prior authorization, and claims at high volume.

  • Validations, retries, and orchestration patterns standardized across processes
  • Improved consistency, throughput, and turnaround across RCM lines of business

Impact: ~40–80% reduction in repetitive effort (varies by workflow)

Stack:PythonSeleniumPandasSQL

Other Work

Academic and research projects

⚙️

Lite Bot Manager (POC)

Lightweight dashboard that validated orchestration, access control, and grouping before full BotVeta rollout.

  • Quick bot launch and basic tracking
  • Informed scheduling and multi-bot UX for the production platform
PythonFastAPIReact
💬

Slack-Based Workflow Automation

Event-driven triggers and monitoring for internal workflows via Slack APIs and webhooks.

  • Real-time execution signals for automation teams
  • API-driven integration with existing bot and pipeline tooling
PythonFastAPISlack API
🧪

Database Development for Microfluidics

Structured database system for microfluidics research data retrieval and analysis.

  • Schema design and data management support
  • Improved organization for downstream analysis
SQLDatabase DesignData Management

Key Achievements

Highlights moved from my résumé into the portfolio so the CV stays scannable—full context lives here.

  • Delivered end-to-end automation systems processing high-volume healthcare workflows with up to 95% reduction in turnaround time.
  • Designed and implemented a multi-tenant orchestration platform (BotVeta) supporting scalable workflow execution and monitoring.
  • Built and deployed AI-powered production systems using AWS Bedrock for document intelligence and IVR automation.
  • Developed distributed, fault-tolerant backend systems with queue-based execution, retry handling, and real-time orchestration.
See featured project write-ups →

BotVeta — Automation Orchestration Platform

Flagship
  • Multi-tenant platform for scheduling, monitoring, and managing distributed automation workflows.
  • Architecture: Orchestration APIs → Scheduler → Bot Runners → Centralized logging & monitoring.
  • Supports queue-based execution, retry handling, failure recovery, and workflow orchestration.
Tech:PythonFastAPI.NETReactNext.jsSQL ServerPostgreSQLAWSDocker

AI-Powered PDF Extraction & Document Intelligence System

  • AI-driven document processing system for extracting structured data from unstructured healthcare PDFs.
  • Implements document splitting, grouping, and validation using rule-based logic with AI confidence scoring.
  • Integrates AWS Bedrock (Claude) for LLM-based extraction.
Tech:PythonFastAPIAWS BedrockLLMs

AI Voice Automation Platform (IVR + Amazon Connect)

  • Real-time IVR automation system for payer workflows.
  • Integrates Amazon Connect with backend orchestration and AI-driven conversational flows.
  • Supports automated navigation, response handling, and dynamic call routing.
Tech:PythonREST APIsAmazon ConnectAWS

RCM Automation Suite (Eligibility — Authorization — Claims)

  • Automation platform for eligibility, authorization, and claims workflows.
  • Implements rule-based validation, retry mechanisms, and workflow orchestration.
  • 40–80% reduction in repetitive effort.
Tech:PythonSeleniumPandasSQL

Slack-Based Workflow Automation System

  • Event-driven system for triggering and monitoring workflows via Slack.
  • Uses APIs and webhooks for real-time execution.
Tech:PythonFastAPISlack API

AI Bots & Utility Systems

  • Built conversational bots (Telegram/WhatsApp) using LLMs.
  • Developed audio transcription pipelines using OpenAI APIs.
Tech:PythonOpenAI APIsFlask

Technologies I Love

Modern tools and frameworks that power my development journey

🐍
Python
🚀
FastAPI
🔷
.NET / C#
⏱️
Hangfire
🎯
Django
🌶️
Flask
Next.js
⚛️
React
Vite
🗃️
SQL Server
🐘
PostgreSQL
🗄️
SQL
🤖
Selenium
☁️
AWS
🧠
Bedrock
📞
Amazon Connect
🐳
Docker
🔌
REST APIs
LLMs
📦
Git
⚙️
CI/CD
📘
TypeScript
📜
JavaScript

Publications

Research contributions and academic achievements

📄

Electrocardiogram Heart Rate Variability Signal Analysis and Data Acquisition: A Novel Approach for Sympathovagal Inference

Vasu, K., Kumar, P., Mane, P. R., Paul, B., Vaishali, K., & Sinha, M. K.

2023 International Conference on Network, Multimedia and Information Technology (NMITCON)

IEEESeptember 2023

Achievements:

🏆Best Presenter Award🏆Best Paper Award

Resume

View my comprehensive resume with detailed experience, skills, and achievements

📄

Karuna Vasu — Resume

Software Engineer — view inline or download the latest PDF résumé.

Quick Summary

Experience

2+ Years

Location

Mangalore, India

Status

Open to Work

Let's Connect

Whether you're looking for a developer or want to discuss automation and AI solutions, I'd love to hear from you.

📍 Mangalore, Karnataka 575003, India