Jaideep Mala

🚀 What I Do

Senior Software Developer at AWS with 8 years of experience building reliable, high-scale systems across Java, Spring, AWS, GCP, MySQL, NoSQL, Android, and modern JavaScript stacks. I design backend platforms, distributed services, cloud architectures, and product workflows that improve latency, reliability, and operating cost. I also build autonomous AI engineering workflows that connect chat, codebases, checks, review, and pull requests into practical delivery systems.

🧠 Skills

Backend Engineering

Java Spring Boot Spring MVC Node.js Python

Cloud & Platform

AWS AWS Lambda AWS Fargate Amazon S3 GCP

Data Systems

MySQL NoSQL MongoDB PostgreSQL Database Design

Distributed Systems

Kafka Redis Microservices High Scale APIs

AI Automated Workflows

Agent Workflows Tool Orchestration Code Automation Review Automation Autonomous PRs Delivery Systems

Product Frontend

React JavaScript Android Vercel Render

⚡ Projects

Restaurant POS SaaS

A browser-based point-of-sale platform for restaurants with live order flow, menu management, kitchen operations, role-based access, configurable billing, and multi-tenant foundations. This is the main product build: Vercel frontend, Render backend, MongoDB Atlas, realtime order updates, and a production domain at pos.jaideepmala.com.

React Node.js MongoDB Socket.IO Vercel Render Multi-tenant SaaS

Live Demo Frontend Backend

Autonomous AI Software Engineering Workflow

I built a working autonomous coding agent around this POS product. From a simple Telegram request, the Node.js workflow creates a tracked job, plans the change, invokes Codex to edit the repository, runs build/test checks, reviews the diff, pushes a branch, opens a GitHub PR, and reports the result back to Telegram.

Telegram request

Job manager

Planner agent

Codex coder

Build & tests

AI review

GitHub PR

Telegram result

1. Telegram command interface

Private Telegram bot accepts natural-language feature requests and returns job progress updates.

2. Job orchestrator

Node.js controller manages lifecycle states, retry attempts, repository targets, artifacts, and logs.

3. Planner and Codex coder

The planner produces implementation intent, then Codex edits the checked-out frontend or backend repo.

4. Automated verification

The workflow runs configured frontend or backend checks and sends failures back into the repair loop.

5. Reviewer agent

A review pass inspects the generated diff for bugs, regressions, missing validation, and test risk.

6. Pull request automation

The bot commits, pushes the branch, opens a GitHub PR, and sends the PR link back to Telegram.

Built as a real product path: frontend → API → database, then extended with staff roles, realtime updates, restaurant-specific settings, and a working autonomous AI PR workflow with planning, Codex code generation, automated checks, review, branch push, and Telegram notifications.

Distributed Job Scheduler

A generic distributed scheduling platform built with Java and Spring Boot. The system accepts one-off and cron jobs through APIs, persists durable job definitions in Postgres, dispatches due executions through Kafka, coordinates worker concurrency with Redis leases, retries failed jobs with backoff, moves exhausted executions to a dead letter queue, and exposes execution state for an operations dashboard.

Java Spring Boot PostgreSQL Kafka Redis Leases Retries DLQ Render Vercel Dashboard

Distributed job scheduler dashboard overview

Fault-Tolerant Scheduling Architecture

The scheduler separates job intake, dispatch, publication, and execution. The API stores job definitions in Postgres. A dispatcher claims due jobs with database locks, creates execution records, and writes outbox events. The outbox publisher sends execution events to Kafka. Worker services consume those events, acquire Redis leases, execute registered job handlers, and update Postgres with success, retry, or dead-letter status.

API

Postgres

Dispatcher

Outbox

Kafka

Workers

Redis lease

DLQ

1. Durable job registry

Postgres stores job definitions, executions, attempts, retry state, and dashboard history.

2. Reliable dispatch

The dispatcher scans due jobs, uses row locks to avoid duplicate scheduling, and writes outbox events.

3. Kafka execution stream

Kafka decouples scheduling from worker execution and allows worker fleets to scale independently.

4. Redis lease control

Workers acquire expiring leases so duplicate Kafka deliveries do not run the same execution concurrently.

5. Retry and backoff

Failures are retried with policy-driven backoff before exhausted executions move to DLQ.

6. Operations dashboard

The dashboard shows job definitions, live executions, attempts, failure reasons, and replayable DLQ entries.

Built as a reusable infrastructure system: Spring Boot services on Render, Vercel dashboard, Postgres as source of truth, Kafka-compatible execution events, Redis lease coordination, retries, dead letter queue, and effectively-once execution semantics.

Hotel Booking Platform

A mobile-first hotel booking journey covering search, hotel discovery, room selection, review, and payment handoff — shaped for high-intent conversion and smooth travel-commerce flows.

Java JavaScript AWS CDN Amazon S3 Travel Commerce