Prasanna Kulal
Coastal sunset over water

Building Modern Web Apps & AI-Powered Systems

Download CV

Full Stack & AI Engineer with hands-on experience building responsive React interfaces and Node.js backends for production AI systems. I work with multiple language models, build AI retrieval tools for coding assistants, and design reliable message processing that prevents data loss under failure.

How I Build

I build systems that keep working when things go wrong — automatic fallbacks, reliable processing, and monitoring across every layer.

Resilient AI Pipelines

When one AI model hits rate limits or goes down, the system automatically switches to another provider. The pipeline keeps running through provider failures without losing any work.

Zero Data Loss

Message queues with proper retry logic, error recovery, and safeguards that ensure no data is lost — even when a worker crashes mid-process.

Full Monitoring

Dashboards, metrics, and structured logging across all services — so issues are caught and diagnosed quickly in production.

Selected Work

Featured Projects

Real projects solving real problems — from AI-powered search to real-time translation and sentiment analytics.

AI Document Search2025

CodexMemory

A document search system that indexes your local files and lets AI coding assistants find relevant context instantly using vector embeddings.

  • Integrated with IDE coding assistants via MCP tools
  • Background processing pipeline with automatic retry and error handling
  • Full monitoring with Prometheus, Grafana, and structured logging
Node.jsMCP SDKBullMQRedisLanceDBPrometheus
Browser Extension2024

MeetLingo

A Chrome extension that translates live captions in real time on Google Meet and Microsoft Teams without breaking page styling.

  • Isolated caption overlay using Shadow DOM
  • Reduced API costs 30–40% with smart caching
  • Rate limiting and automatic retry logic
Chrome MV3JavaScriptShadow DOMDeepL APIGoogle Cloud
Analytics Platform2025

PRISM

An analytics platform that tracks brand sentiment across YouTube, X, Reddit, and web sources through a multi-stage AI pipeline.

  • Automatic switching between AI models when one fails
  • Background job scheduling for crawling and batch processing
  • Real-time dashboard with TypeScript, Zustand, and TanStack Query
ReactNode.jsExpressPostgreSQLPrismapg-bossTailwindCSS
Background

Experience

Building production AI systems, analytics dashboards, and cloud infrastructure.

Associate Software Engineer — Zynthora AI

Sep 2025 – Present

Full-stack and AI engineer working on cloud infrastructure and production AI workflows. Building backends that handle model failures gracefully, process messages reliably, and provide real-time visibility into system health.

AI Pipelines

Multi-step AI processing pipelines with error recovery and automatic model switching.

Backend Systems

Redesigned message processing for zero data loss, with Redis caching and proper retry logic.

Monitoring

Full monitoring and structured logging across all services with dashboards for quick diagnosis.

Toolkit

Technical Stack

Languages & Core

  • TypeScript & JavaScript
  • Python
  • Java (Core & OOP)
  • SQL

AI & Tools

  • Gemini, Grok, OpenAI
  • LangChain
  • RAG Pipelines & Vector Search
  • Model Context Protocol (MCP)

Backend & Data

  • Node.js & Express
  • RabbitMQ & BullMQ
  • Redis & pg-boss
  • PostgreSQL & Prisma

Infra & Frontend

  • AWS (EC2, Secrets, CloudWatch)
  • Docker & Docker Compose
  • React, Next.js, TailwindCSS
  • Prometheus, Grafana & Loki
Get in Touch

Let's connect....

Send me a message or reach out directly, I'll get back to you as soon as I can

Your details are used only to reply to your message.