senior software engineer

Swarna Avvaru

Front End · Full-Stack · Cloud-Native

8+ years designing, building, deploying, and supporting production systems across Angular, React, .NET, Python, and AWS. More recently, I've expanded into AI-powered developer tools — multi-agent applications and AI-assisted engineering workflows built on modern LLMs.

View projects Email me
profile.json
experience
8+ yrs
frontend
angular · react · typescript
backend
.net core · python · REST APIs
cloud
aws lambda · messaging · monitoring
ai_workflow
self hosting llm · langgraph · rag
status
available

stack manifest

Engineering across the modern software stack.

Building production-ready applications across frontend, backend, cloud, and AI — from design and development to deployment, monitoring, and continuous improvement.

frontend

  • Angular
  • React
  • TypeScript
    • li>Javascript
  • RxJS
  • NgRx
  • Redux
  • Kendo UI

backend

  • C#
  • .NET Core
  • ASP.NET Core Web API
  • Python
  • FastAPI
  • Entity Framework
  • REST
  • Microservices

cloud

  • AWS Lambda
  • API Gateway
  • SNS
  • SQS
  • S3
  • CloudWatch
  • Azure DevOps
  • GitHub Actions
  • Datadog

ai

  • OpenAI/Claude APIs
  • LangGraph
  • Prompt Engineering
  • RAG Workflows

selected builds

Practical builds that show range.

live featured · ai project

AI Incident Copilot

Claude-powered production incident assistant. Feed it symptoms, logs, and runbooks — it returns a structured investigation brief: probable cause, evidence, next steps, and stakeholder-ready updates.

  • React + TypeScript frontend, Python FastAPI backend
  • Local RAG over uploaded runbooks and built-in troubleshooting guides
  • Backend-only API key handling, usage limits, mock mode — deployed on Vercel + Render
Live demo ↗ Source code ↗
live new · multi-agent

ReviewPilot

Multi-agent PR review copilot. Specialist agents inspect a diff for security, performance, architecture, and test coverage — a Judge agent merges duplicate findings and filters speculative ones before anything reaches a reviewer.

  • LangGraph pipeline: four specialist agents + a Claude-powered Judge for dedupe and severity calibration
  • PR mode (GitHub API diff ingestion) and a Playground mode for reviewing pasted code with no repo access
  • FastAPI + SQLite/Postgres backend, React frontend — deployed on Vercel + Render
Live demo ↗ Source code ↗
Modal AI Chat — self-hosted Qwen chat interface
live

Modal AI Chat

Self-hosted Qwen chat app using Next.js, Modal GPU inference, vLLM streaming, rate limits, and server-side API protection.

Live demo ↗ Source code ↗

learning & writing

Continuous growth.

Certifications, courses, and published writing.

let's connect

Open to Front End, Full Stack, and AI Engineer opportunities.

I'm excited to join teams building reliable products with modern web and cloud technologies — whether it's developing new features, improving existing systems, or building AI-powered experiences.

swarnakishoree@gmail.com