Building production-grade distributed systems with automated AWS deployments, achieving sub-5ms response times at 1K+ concurrent users

🔍 Actively seeking: Remote backend engineering positions
🚀 Specialization: Python backend + DevOps automation + Distributed systems
📍 Location: Dhaka, Bangladesh (Open to worldwide remote)
💬 Ask me about: FastAPI, System Design, AWS Infrastructure

I don't just write backend code—I architect complete production systems with full automation from infrastructure to deployment:

✅ Infrastructure as Code Expert - Automated AWS deployments managing 11+ EC2 instances with Pulumi & Ansible
✅ Performance Engineering - Optimized systems achieving sub-5ms response times with 1K+ concurrent users
✅ DevOps Automation - Zero-touch deployments with CI/CD, containerization, and orchestration
✅ Distributed Systems - Built fault-tolerant architectures with auto-scaling, load balancing, and high availability
✅ Technical Writing - Published articles explaining complex architectures in simple words

🎯 6+ Production-Ready Applications Built
⚡ Sub-5ms API Response Times Achieved
🖥️ 11+ AWS EC2 Instances Automated
📦 1K+ Concurrent Users Supported
🔄 Container Orchestration Systems Designed
🧪 500+ DSA Problems Solved
📖 200K+ Technical Blog Readers
🎥 40+ Educational Videos Created

High Complexity - Decoupled Microservices Architecture

High-performance URL shortener with three independent services, dual database strategy, and production K3s deployment

• Architected decoupled microservices separating create_service (write-heavy), redirect_service (95% read traffic), and worker_service (Celery tasks) with independent scaling via docker-compose-decoupled.yml.
• Implemented Redis-first caching with MongoDB fallback and Nginx proxy routing, targeting sub-5ms redirect latency for optimal user experience.
• Built repository pattern with abstract base classes for PostgreSQL and MongoDB, centralized error handling, and shared common/ library for clean data access across services.
• Optimized PostgreSQL operations using atomic key acquisition with SELECT FOR UPDATE SKIP LOCKED for race-free distributed key allocation and parameterized bulk inserts for 100K+ keys.
• Implemented comprehensive observability using OpenTelemetry with B3 propagation, automatic FastAPI/DB instrumentation, OTLP export to Tempo for distributed tracing and Grafana visualization.
• Engineered production-grade resilience with PgBouncer connection pooling (53% reduction in overhead), circuit breaker pattern preventing cascade failures, exponential backoff retries, and database timeout protection.
• Deployed production K3s cluster on AWS using Pulumi IaC and Ansible with path-based Nginx routing, per-service rate limiting, and CI/CD pipeline via GitHub Actions.
• Implemented intelligent key pre-population using Celery workers maintaining pool of unused keys for instant URL creation without database latency.
• Built comprehensive testing infrastructure with multi-database mocking (SQLite, mongomock, fakeredis), async pytest framework, httpx API client testing, and isolated test environments.
• Automated AWS infrastructure with VPC setup, security groups, bastion host access, and monitoring via Celery Flower dashboard.

Technical Deep Dive: Read my Medium articles

Tech Stack: FastAPI Redis PostgreSQL MongoDB Celery Nginx Docker K3s Pulumi Ansible AWS OpenTelemetry Tempo Grafana PgBouncer pytest httpx GitHub Actions

Key Learnings:

• Microservices decoupling for independent scaling
• Repository pattern for clean data access
• Circuit breaker pattern for fault tolerance
• Multi-database testing strategies
• Infrastructure as Code best practices

Most Complex Infrastructure Project - ML-Enhanced Event-Driven Architecture

Production-grade autoscaling system for K3s clusters with 4-layer intelligent scaling architecture, ML-based predictive scaling, and multi-AZ high availability

• Architected 4-layer autoscaling system: (1) Data Collection for ML training, (2) Time-Aware Scaling with peak/off-peak thresholds (85%/60% vs 60%/40%), (3) Flash Sale Detection with emergency response to CPU spikes >30% in 2 minutes, (4) Predictive Scaling using Prophet models forecasting CPU 15 minutes ahead.
• Implemented ML training pipeline with Kubernetes CronJob for automated weekly model retraining, feature engineering (temporal cyclical encoding, lag features, rolling statistics), time-series cross-validation, and backtesting with MAE/RMSE metrics tracking.
• Built event-driven Lambda architecture with four specialized functions (Decision, Scale-Up, Scale-Down, Cleanup) orchestrated through EventBridge for fault tolerance, crash recovery via Write-Ahead Log (WAL), and distributed locking with 200s timeout.
• Designed multi-AZ high availability with round-robin worker distribution across 3 availability zones (ap-southeast-1a/b/c), single NAT Gateway optimization, and LIFO scale-down maintaining natural distribution balance.
• Implemented multi-layer idempotency including bootstrap verification, cooldown checks (scale-up: 300s, scale-down: 900s), pending instance detection, and automatic stale flag cleanup to prevent duplicate scaling operations.
• Integrated comprehensive observability with 17 CloudWatch alarms (CRITICAL/WARNING severity), Prometheus health graceful degradation (conservative defaults when unavailable), and fixed LogGroups for stable dashboard references.
• Engineered spot instance support with automatic On-Demand fallback when spot capacity unavailable (InsufficientInstanceCapacity, SpotInstanceCapacityNotAvailable, MaxSpotInstanceCountExceeded), graceful 2-minute interruption handling, and proper node cleanup.

Tech Stack: AWS Lambda EventBridge DynamoDB EC2 K3s Prometheus CloudWatch Prophet Kubernetes CronJob SSM Secrets Manager S3 Python 3.11 Pulumi Ansible kubectl Node Exporter

Key Learnings:

• Layered autoscaling architecture combining reactive (time-aware, flash sale) and proactive (ML predictive) scaling
• Event-driven architecture patterns with Lambda chaining via EventBridge
• Distributed systems state management with DynamoDB and WAL patterns
• ML pipeline deployment with automated retraining and model versioning
• Multi-AZ infrastructure design with cost optimization (single NAT, spot instances)
• Kubernetes cluster operations including node lifecycle, pod draining, and CronJob scheduling

High Complexity - Media Processing Pipeline

Full-Stack advanced video streaming solution with adaptive bitrate technology

• Engineered a secure and scalable video platform with a Django REST API and a React/TypeScript frontend, architected for high-performance adaptive streaming.
• Implemented a robust security model, using dj-rest-auth for token-based authentication and a protected media workflow (via Nginx X-Accel-Redirect) to ensure only authorized users can access streaming content.
• Built an asynchronous video processing pipeline using Celery, Redis, and FFMPEG to transcode videos for DASH playback, ensuring a smooth, low-latency user experience.
• Automated the entire cloud workflow, from provisioning AWS S3 infrastructure with Pulumi and configuring servers with Ansible, to deploying the Docker-containerized application via GitHub Actions.

Tech Stack: Django React Celery Redis PostgreSQL FFMPEG DASH AWS S3 Nginx Docker Pulumi Ansible

Medium-High Complexity - Worker Orchestration

Scalable job processing system with advanced features

• Developed a distributed job queue system using FastAPI and Redis to manage asynchronous tasks with priority-based queuing and automatic worker scaling.
• Implemented a real-time monitoring dashboard with Jinja2 templates to provide visibility into job status, queue metrics, and worker activity.
• Engineered an automatic worker scaling mechanism based on job load and worker availability, using Docker Swarm to dynamically adjust resources.
• Created a comprehensive error handling and fault tolerance system, including automatic retries for failed jobs and a dead-letter queue for unrecoverable tasks.
• Designed a job dependency feature to ensure complex workflows are executed in the correct order, improving system reliability.
• Containerized all services (API, Worker, Monitor) using Docker for consistent deployment and simplified management.

Tech Stack: FastAPI Redis Docker Swarm Jinja2

Medium Complexity - Full-Stack Application

Full-stack financial management application for tracking installments and payments

• Backend: High-performance API built with FastAPI, using SQLAlchemy for ORM with a PostgreSQL database.
• Frontend: Modern and responsive UI built with React, TypeScript, and Vite, styled with Tailwind CSS and Shadcn UI.
• Asynchronous Tasks: Celery and Redis manage background jobs like sending OTP and due date notification emails.
• Authentication: Secure JWT-based authentication with role-based access for customers and admins.
• Data Management: Alembic handles database schema migrations, and TanStack Query manages server state on the frontend.
• DevOps: Fully containerized with Docker and Docker Compose for reproducible development and deployment environments.

Tech Stack: FastAPI React TypeScript PostgreSQL SQLAlchemy Redis Celery Docker Tailwind CSS Shadcn UI Alembic

Medium Complexity - Async Communication

Real-time notification system for multiple channels

• Modern Backend: Built with Python and FastAPI for high-performance, asynchronous API endpoints.
• Multi-Channel Delivery: Supports sending notifications through various channels like Email, SMS, and Push Notifications.
• Asynchronous & Scalable: Leverages Celery and RabbitMQ for background task processing, ensuring the system can handle high-volume loads without blocking.
• Robust Data Storage: Uses PostgreSQL for reliable data persistence, managed with Alembic for smooth database migrations.
• Containerized Environment: Fully containerized with Docker and Docker Compose for consistent development, testing, and deployment.
• Comprehensive Testing: Includes a full suite of tests using pytest to ensure code quality and reliability.

Tech Stack: FastAPI Celery PostgreSQL RabbitMQ Redis Alembic SQLAlchemy Docker Pytest

Medium Complexity - HA Architecture

Enterprise-grade Todo application with AWS infrastructure

• Engineered full-stack application with FastAPI backend and React frontend
• Implemented Infrastructure as Code using Pulumi for AWS resource management
• Designed fault-tolerant architecture with load balancing across multiple AZs
• Built PostgreSQL replication system with automated backup/recovery
• Integrated Redis Sentinel for high availability caching

Tech Stack: FastAPI React AWS EC2 PostgreSQL Redis Sentinel Nginx Docker

June 2024 - August 2024 | Dhaka, Bangladesh

🎯 Delivered measurable business impact:

• Designed RBAC dashboard for 200+ users with real-time analytics
• Automated 40% of manual processes through intelligent workflows
• Built production-ready meal scheduling system with cron jobs

Bachelor of Science in Computer Science & Engineering
Daffodil International University | September 2017 - December 2022

• Building a Scalable URL Shortener: System Design to Production
  • Complete architectural breakdown with Infrastructure as Code
  • 100+ views, featured in system design discussions

• 200,000+ readers on Quora with tech insights in Bengali
• Nearly 200 followers engaging with technology content
• 40+ instructional videos on YouTube bridging Bengali tech education gap

• 500+ Problems Solved across multiple platforms
• Active on: BeeCrowd, LightOJ, HackerRank, LeetCode
• Contest Achievements:
  • DIU Take-Off Programming Contest (Ranked 6th out of 300 participants)
  • Multiple university-level programming contest participations

• 🐳 Kubernetes - Container orchestration at scale

I'm actively seeking opportunities to work on:

• 🏗️ Distributed systems requiring high availability and fault tolerance
• ☁️ Cloud-native applications with automated infrastructure
• 🔄 Microservices architectures with proper observability
• 📚 Open-source projects where I can contribute infrastructure expertise

Looking for a backend engineer who can:

• ✅ Design scalable distributed systems
• ✅ Automate infrastructure from scratch
• ✅ Write clean, testable, maintainable code
• ✅ Document complex architectures clearly

Let's build something amazing together!

• 📧 Email: kaziiriad@gmail.com
• 📱 Phone: +880 1683152495
• 💼 LinkedIn: Sultan Mahmud
• 📝 Medium: @kazisultanmahmud
• 📺 YouTube: I.T. Darshonik