Hologram Matrix Matrix Grid

Hi! I'm

Jagadeeswar Jonnadula

Hey there! I'm a backend-focused software engineer who loves building stateless, scalable, cloud-native architectures, distributed systems, and AI-powered automation pipelines.

Resume

About Me

I am fundamentally a backend-focused software engineer with an implementation-first mindset. I solve system-level problems over UI-focused tasks, with a clear bias toward infrastructure logic, data flow design, event-driven systems, and cloud-native architectures. I thrive on handling real-world production constraints — retries, failures, idempotency, and distributed consistency.


Core Engineering Strengths


Backend & Cloud Architecture

Backend & Cloud-Native Systems

Designing stateless, scalable, cloud-native architectures on AWS and GCP. Highly experienced with microservices, serverless computing (Lambda, API Gateway, EventBridge), and optimizing APIs and data pipelines for heavy production loads.

Distributed Concurrency

Distributed Concurrency

Handling horizontal scaling, workload distribution, and concurrency. Skilled in lease-based concurrency patterns with Firestore atomic transactions to process concurrent workflows without failures.

Reliable Payments & Billing

Reliable Payments & Billing

End-to-end payment and subscription systems integration using Razorpay, PayPal, and Apple Pay. Engineering robust transaction lifecycle handlers, webhook verification, double-charge prevention, and idempotent transactions.

Production AI Integration

Production AI Integration

Integrating AI as a reliable system component. Building automated multi-agent evaluation pipelines (Vertex AI, Gemini, Claude) with context-aware conversation history, prompt optimizations, and RAG architectures.



Education & Certifications


Education VNRVJIET

Education — VNRVJIET

VNR Vignana Jyothi Institute of Engineering and Technology
B.Tech — Computer Science and Engineering (Data Science)
2023 – 2027 | Hyderabad, India
CGPA: 9.1

Certifications

Professional Certifications

AWS Certified Cloud Practitioner Essentials
Amazon Web Services | March 2025

SQL (Intermediate)
HackerRank | May 2024

GitHub Streak

GitHub Status

GitHub Languages GitHub Stats

Snake animation

Work Experience

SDE Intern

Realityrift Innovations (Remote)
June 2025 – February 2026

Stack: TypeScript, AWS (Lambda, DynamoDB, EventBridge, CloudWatch, S3, SES), GCP (Vertex AI, Firebase), Next.js, Git

  • Maintained and improved production stability across HelloAria (helloaria.in) — a serverless AI productivity assistant with deep Google Workspace integration running on a 35+ microservice architecture.
  • Subscription Billing & Auth: Integrated end-to-end global subscription billing with Razorpay, PayPal, and Apple Pay. Resolved critical Firebase authentication and 2FA issues affecting sign-in flows.
  • Reminder Delivery System: Built a reminder delivery system on AWS EventBridge + Lambda with configurable retry logic across user-preferred channels — WhatsApp, email, and dashboards.
  • Vertex AI Pipelines: Constructed multimodal pipelines (image, audio, document) using Vertex AI/Google Gemini. Added invisible context injection for improved multi-turn conversation awareness. Built a meeting document generator that processes audio into structured Google Docs summaries (MoM generator).
  • Admin Observability: Built internal admin analytics dashboards using CloudWatch metrics. Implemented Apple OAuth sign-in, dashboard chat interface, and REST APIs serving both web and native clients.
  • Engineered and optimized multimodal AI prompts for conversation summarization, image analysis, and automated actions — significantly improving response accuracy and task-completion reliability.

Freelance Developer

Absolute Laddoos
May 2025 – June 2025

Stack: Next.js 15, Razorpay, Firebase, GCS, TypeScript

  • Designed and deployed a full-stack e-commerce platform with Next.js 15 and Firebase, ensuring 99.9% transaction reliability.
  • Payment Engineering: Implemented complete payment flows with Razorpay. Handled failed transactions and double-charges using Firebase transactions; built SHA-256 duplicate-order prevention and race-condition-safe bulk order processing.
  • Admin Dashboard & CDN: Built a comprehensive admin panel with bulk operations, coupon engine, geo-based pricing, and background order confirmation. Optimized image assets using GCS with Cloud CDN.

Developer

SportsFest Registration
Quasi-Work

Stack: React.js, Node.js, Express.js, Nodemailer, Razorpay, Firebase

  • Maintained a full-stack registration platform for 200+ participants with Firebase Realtime Database and Authentication.
  • Implemented Razorpay webhook handling and backend refactoring, ensuring ₹3.8L+ revenue was reliably verified.
  • Automated email notifications via Nodemailer and optimized form handling for a smoother, low-friction registration UX.

Key Projects

VNR Reports

VNR Reports

A production-grade institutional reporting platform actively used by 20–25+ departments at VNRVJIET. Built dynamic JSONB-driven forms with Role-Based Access Control (RBAC) and GCS file uploads. Developed a server-side DOCX generator that compiles rich text and JSON data into formatted 50+ page reports with real-time tracking via Server-Sent Events (SSE).

PostgreSQL RBAC GCS SSE TypeScript
Agentic Evaluator

Agentic Evaluator Engine

A fully automated multi-agent evaluation pipeline that scores system design diagrams using Descriptor, Evaluator, and MCQ Generator agents built on Vertex AI. Processed 70+ concurrent submissions at ~70s per evaluation, distributed across stateless workers with Firestore atomic transactions for lease-based concurrency.

Vertex AI Firestore Concurrency Leasing Python
Trovo Rewards

Trovo — Credit Card Rewards Platform

Full Stack Stack: React 18, TypeScript, Vite, Tailwind, TanStack Query, Express.js, Drizzle ORM, Supabase (PostgreSQL), Zod, Passport.js

  • Backend (Rebuilt from Scratch): Designed modular Rest APIs with separate domain modules, JWT-based stateless auth, Drizzle ORM, and Drizzle-Zod schema-driven runtime validation.
  • Optimization: Built aggregated dashboard endpoints utilizing Promise.all to fan out parallel SQL queries in a single round trip, lowering dashboard cold-starts.
  • Frontend (Refactored): Cleaned up a hackathon SPA into a production-shaped app with 8 routes, 46 UI components, TanStack state management, client CSV exports, points trend charts, and an AI chat assistant.
  • Scale: Fully multi-tenant database supporting 8 major Indian banks, 40+ card products, and 30+ redemption pathways.
Emergency Detection

Emergency Vehicle Detection System

Engineered a real-time computer vision pipeline to detect emergency vehicles, achieving 0.780 mAP@0.5 for ambulances (105% improvement over baseline). Implemented frame-skipping achieving 30-50+ FPS on standard hardware, with an integrated traffic light simulator that prioritizes emergency vehicles automatically.

YOLOv8n OpenCV PyTorch Streamlit
OceanQuery

OceanQuery

Full-stack web application for querying and visualizing complex oceanographic data using natural language. Backed by PostgreSQL with pgvector for semantic retrieval, rendering real-time interactive charts and geospatial maps.

React Node.js pgvector PostgreSQL
InfraOps

InfraOps

A generative AI-powered SaaS platform for automated ticket classification in residential communities. Built with Google Gemini AI to classify maintenance categories with 90-95%+ accuracy, reducing manual technician assignments by 70-75%. Features image-based issue recognition and spam/duplicate detection.

Google Gemini SaaS Architecture Firebase React
Univo

Univo

Built during Webathon 2025 under a constrained 16-hour timeframe. Campus event management platform enabling students to discover, register for, and track events with role-based access, Razorpay ticketing, Nodemailer alerts, and gamification. Won 1st Place out of 50+ teams.

React.js Razorpay API Nodemailer Firebase