Medium

Abul Kalam Azad

Pinned

In

Level Up Coding

by

Abul Kalam Azad

·

Oct 28

The Vibe Coding Hangover: What Happens After AI Writes Your Codebase

AI can build your MVP faster than ever. But without architectural ownership, the real cost shows up in production.

The Vibe Coding Hangover: What Happens After AI Writes Your Codebase

Pinned

In

Level Up Coding

by

Abul Kalam Azad

·

Sep 25

The Vibe Coding Trap: Why AI-Generated MVPs Are Quietly Bankrupting Early-Stage Startups

A founder friend recently showed me a full-stack SaaS he built over a single weekend using Cursor. Two weeks later, his database completely…

$Conceptual editorial 3D illustration of a sleek dark computer screen with glowing neon metrics and subtle hairline fractures, symbolizing fragile AI-generated software architecture.$

$Conceptual editorial 3D illustration of a sleek dark computer screen with glowing neon metrics and subtle hairline fractures, symbolizing fragile AI-generated software architecture.$

Pinned

In

Level Up Coding

by

Abul Kalam Azad

·

Sep 23

How to Scale PostgreSQL to 100,000 Writes Per Second Without Sharding

When engineering teams hit write-throughput walls in PostgreSQL, the default architectural reflex is often distributed sharding (Citus…

High-performance PostgreSQL database scaling architecture showing massive parallel write ingestion into a database engine.

Abul Kalam Azad

·

8h ago

uWhy AI Agents Fail in Production: The Engineering Problems Beyond the LLM

Building reliable AI agents requires more than better models. It requires orchestration, durable state, idempotency, observability…

A glowing AI agent module connected to complex industrial backend infrastructure, surrounded by tangled cables, warning indicators, and production system components.

Abul Kalam Azad

·

1d ago

Stop Just Using AI APIs. Learn to Engineer AI Systems in 2027

A practical 12-month roadmap to mastering Python, machine learning, LLMs, RAG, and AI agents — and building reliable, production-ready AI…

A visual representation of the journey from using AI APIs to engineering production-ready AI systems, featuring artificial intelligence, machine learning, large language models, and connected backend infrastructure.

Abul Kalam Azad

·

4d ago

Go Inside Next.js: When Developer Convenience Hides Architectural Complexity

Why adding compiled languages to a React meta-framework raises important questions about runtime isolation, independent scaling, and…

A futuristic illustration showing Next.js and Go runtimes connected through an integration boundary, highlighting architectural complexity, runtime isolation, independent scaling, vendor lock-in, and operational challenges.

In

Level Up Coding

by

Abul Kalam Azad

·

Oct 26

The Dirty Truth About Autonomous AI Agents: Why 100-Line Demos Break in Production

An AI agent that can complete a task is not necessarily a system you can trust. Reliability comes from engineering, not autonomy alone.

A 3D abstract visualization of an autonomous AI agent workflow bounded within structured architectural state machines and security guardrails on a dark background.

In

Level Up Coding

by

Abul Kalam Azad

·

Oct 21

The Architectural Guide to Distributed Data Consistency: Why Dual-Writes Break Microservices and…

Handling partial failure across PostgreSQL, Apache Kafka, and external APIs using the Transactional Outbox Pattern and deterministic…

A minimalist 3D abstract illustration of distributed data consistency showing synchronized database storage and real-time event streams connected via glowing pipelines on a dark background.

Abul Kalam Azad

·

Oct 18

Coding Isn’t Dead — Typing Is: Why AI Can Write 10,000 Lines but Can’t Architect a Distributed…

AI makes code generation cheap. Production reliability still depends on concurrency control, system invariants, and rigorous verification.

Coding Isn’t Dead — Typing Is: Why AI Can Write 10,000 Lines but Can’t Architect a Distributed…

Abul Kalam Azad

·

Oct 14

Beyond the Wrapper: How to Build an AI Micro-SaaS with Actual 80% Gross Margins

Semantic caching, dynamic model routing, and defensive unit economics for engineers who want profit over hype.

A 3D isometric diagram illustrating an intelligent AI request routing pipeline with semantic caching layers and cloud infrastructure optimization on a dark background.

Abul Kalam Azad

Abul Kalam Azad

Senior Software Engineer @ Arogga | Leading backend architecture for e-commerce and health-tech | Writing about scalable systems, and real production lessons

Following

Help

Status

About

Careers

Press

Blog

Store

Privacy

Rules

Terms

Text to speech