Md. Rakib Uddin

About Me

Production Backends, SaaS Platforms & AI-Powered Features

Senior software engineer with 8+ years building production PHP and Laravel systems — multi-tenant SaaS platforms with subscription billing, REST APIs, marketplaces, POS and LMS products, and high-traffic publishing systems that run every day for real businesses.

My work centres on backend architecture: modelling the domain properly, designing API contracts that survive change, moving heavy work into queues and events, and using caching and query optimisation to keep response times predictable as data grows. I choose between a well-structured modular monolith and service separation based on team size and real scaling pressure — not fashion.

Over the last cycle I've extended that foundation into AI application engineering: integrating LLM APIs into Laravel applications with structured outputs and tool calling, building RAG pipelines over documents and databases with embeddings and vector search, and automating operational workflows with n8n and webhooks. I build applications on top of existing models and infrastructure — I'm an AI application engineer, not an ML researcher, and the profile below is levelled honestly.

On the delivery side I ship the whole path: Dockerised environments, CI/CD with GitHub Actions, Linux servers, queue workers and scheduled jobs, logging and error tracking. Published commercial products on CodeCanyon and Codester, and led teams from requirements through launch and post-release monitoring.

Architecture-Led Engineering

SOLID, design patterns, service layers and modular boundaries — structure chosen for the problem, not for the résumé.

Security & Reliability First

OWASP Top 10 mitigation, RBAC, tenant isolation, and a code audit before every release. Logging and monitoring after it.

Product Engineering Mindset

Turns business requirements into APIs, data models and roadmaps. Leads sprints, reviews PRs, mentors developers.

Ships to Production

Docker, GitHub Actions CI/CD, Linux, queue workers and cron. Commercial products and live platforms serving real users.

AI Engineering

Building AI Into Real Applications

LLM Application Integration

LLM APIs wired into Laravel services behind clean interfaces, so models stay swappable
Structured outputs and JSON schemas for responses the application can actually rely on
Function / tool calling that lets a model trigger real business logic under validation
Prompt engineering with versioned prompts, evaluation, and guardrails
Token-cost, rate-limit, retry and fallback handling in queued background jobs

RAG & Semantic Search

Retrieval-augmented generation over documents, product catalogues and internal databases
Chunking, embedding generation and storage in PostgreSQL with pgvector
Similarity search combined with SQL filters and permissions, so retrieval respects tenancy
Hybrid keyword + vector retrieval for better recall on domain vocabulary
Answers grounded in cited sources rather than free-running model output

AI Assistants & Agents

In-app assistants that answer from a company's own data instead of generic knowledge
Agentic flows where a model plans, calls tools, and reports back with a bounded loop
Tool definitions mapped to existing Laravel actions with authorisation kept server-side
MCP-based integrations to expose application capabilities to AI clients
Human-in-the-loop checkpoints for anything that writes, spends or sends

AI Workflow Automation

n8n workflows connecting Laravel apps to CRMs, inboxes, sheets and messaging
Webhook and event-driven pipelines that classify, summarise, extract and route content
Document and email processing: extract structured fields, validate, then persist
Scheduled AI enrichment jobs running on queue workers, not in the request cycle
Automation designed with idempotency and replay in mind

AI-Powered SaaS Features

Natural-language search and filtering across tenant data
Summarisation, tagging and categorisation for content-heavy platforms
Draft generation for descriptions, replies and reports, always editable by a human
Per-tenant usage metering, quotas and cost control for AI features
Graceful degradation so the product still works when a provider is down

AI-Assisted Development

AI-assisted coding, refactoring and test generation to compress delivery time
Documentation and changelog generation from real code, kept in review
Debugging and architecture exploration as a second opinion, never as the decision
Code review assistance layered on top of human review, not replacing it
Every generated line is read, tested and owned before it reaches production

Positioning, stated plainly: I build applications on top of existing AI models and infrastructure — integration, retrieval, tooling, cost control and production reliability. I do not train or research models, and I don't present myself as an ML scientist. See the proficiency matrix for where each AI capability sits today.

What I Offer

Services

Multi-Tenant SaaS Development

End-to-end Laravel SaaS: tenancy and data isolation, subscriptions and billing, roles and permissions, admin dashboards, customer portals, white-label configuration, and notifications.

API Platform & Integrations

Versioned, documented REST APIs (and GraphQL where it fits) with Sanctum/Passport auth, rate limiting and backward-compatible contracts. Payment gateways, webhooks and third-party services integrated cleanly.

AI Feature Development

Add LLM-powered capability to an existing product: assistants grounded in your own data, RAG and semantic search with embeddings and pgvector, tool calling into your business logic, plus usage metering and cost control.

Performance & Scalability

Query tuning and indexing, Redis caching layers, N+1 elimination, async queues and horizontal scaling — making slow systems fast without a rewrite. Plus OWASP Top 10 hardening, RBAC and pre-release security audits.

DevOps & Production Delivery

Dockerised environments, GitHub Actions CI/CD, Linux server setup, zero-downtime deploys, queue workers and schedulers, environment and secret configuration, logging, monitoring and error tracking.

Legacy Modernisation & Tech Leadership

Migrating dated PHP/CodeIgniter codebases to modern Laravel, refactoring monoliths into modular architectures without downtime — and leading teams through sprint planning, architecture decisions, PR review and mentoring.

Technical Skills

Engineering Capabilities by Domain

Backend Engineering

PHP 8, Laravel, REST APIs, GraphQL, API architecture, authentication & authorization, validation, queues, events, caching, multi-tenancy, SaaS architecture

Frontend Engineering

Laravel Blade (primary), HTML5/CSS3, JavaScript, Bootstrap, Vue.js, jQuery, React.js, TypeScript, API-driven and AI-powered interfaces

AI Engineering

LLM APIs, prompt engineering, structured outputs, function/tool calling, embeddings, semantic search, RAG, vector search, AI agents, MCP

Data & Storage

MySQL, PostgreSQL, Redis, pgvector, vector databases, data modelling, query optimisation, indexing strategy

DevOps & Cloud

Docker, Git, GitHub Actions, CI/CD, Linux, Nginx/Apache, DigitalOcean, AWS fundamentals, Kubernetes fundamentals, monitoring & logging

Automation

n8n, workflow automation, webhooks, event-driven automation, third-party API integrations, AI workflow automation

Architecture

Clean Architecture, SOLID, design patterns, modular monolith, microservices where justified, event-driven architecture, API-first design, observability

SaaS Product Capability

Multi-tenancy, subscriptions & billing, payment gateways, roles & permissions, notifications, document management, white-label config, admin dashboards, customer portals

Security

OWASP Top 10, CSRF/XSS/SQL injection prevention, secure sessions, HTTPS enforcement, RBAC, tenant data isolation, secret management

Performance

Query tuning, Redis caching, N+1 prevention, eager loading, async queue processing, horizontal scalability planning

Debugging & Observability

Xdebug, Laravel Telescope, structured log analysis, error tracking, root cause identification, post-release monitoring

Product Engineering

Requirement analysis, technical trade-off evaluation, API & schema design, MVP delivery, legacy modernisation, responsible AI-assisted development

Architecture & Engineering Practice

How I Build Systems

Architecture Chosen for the Problem

Starts from domain modelling and separation of concerns, then picks the structure: a well-bounded modular monolith for most products, service separation or microservices only where team scale or load actually justifies the operational cost. SOLID and design patterns applied as tools, not decoration.

Security & Tenant Isolation

OWASP Top 10 mitigation as routine: parameterised queries, CSRF/XSS/SQL injection prevention, secure sessions, HTTPS enforcement, RBAC with Spatie Laravel Permission, and strict tenant scoping so one customer can never read another's data. Code audit before every release.

Asynchronous & Event-Driven Design

Keeps the request cycle thin: heavy work moves to queues and background jobs, side effects run through events and listeners, scheduled tasks handle recurring load. Idempotent handlers and retry policies so a failed job is recoverable rather than a data problem.

API-First Development

Designs the contract before the implementation: versioned REST endpoints, consistent resources and error shapes, Sanctum/Passport auth, rate limiting, and backward compatibility so mobile apps and integrations don't break on deploy. GraphQL where a client genuinely needs flexible querying.

Performance, Data & Observability

Query tuning and indexing, Redis caching, N+1 elimination and eager loading keep response times predictable as data grows. Xdebug, Laravel Telescope, structured logs and error tracking turn production issues into root causes — and long-term fixes rather than patches.

Requirements to Production Ownership

Translates business requirements into architecture, APIs and data models, weighs trade-offs between speed, maintainability, scalability and cost, then owns delivery: sprint planning, code review, mentoring, deployment and post-release monitoring. Uses AI-assisted development to move faster while keeping human review and engineering judgement in charge.

Honest Levelling

Technology Proficiency Matrix

Every technology on this page is placed in one of four tiers. Nothing is listed because it is popular — if it is early for me, it says so.

Expert

Years of professional production work. I own architecture decisions here.

PHP 8LaravelREST API DesignEloquent & Query Optimisation MySQLQueues & JobsEvents & ListenersCaching Strategy RBAC / Spatie PermissionMulti-Tenant SaaSOOP & SOLID Service Layer / RepositoryMVC & Design PatternsOWASP Top 10 Laravel BladeGitLinux / Nginx DeploymentLegacy PHP Modernisation

Proficient

Used confidently in real delivery, with less depth than the expert tier.

Vue.jsJavaScriptRedisSanctum / Passport Payment Gateway IntegrationThird-Party API IntegrationWebhooks DockerComposer / npmBootstrapjQuery Laravel Telescope / XdebugDigitalOceanApache LLM API IntegrationPrompt EngineeringStructured Outputs AI-Assisted Development

Working Knowledge

Built with these; productive from day one, still deepening.

PostgreSQLTypeScriptReact.jsGraphQL GitHub ActionsCI/CDFunction / Tool CallingEmbeddings Semantic SearchRAG PipelinespgvectorVector Databases n8nWorkflow AutomationAWSMonitoring & Logging

Currently Developing

Active learning and side-project work — improving, not yet claiming depth.

Agentic AIAI AgentsMCPKubernetes Fundamentals Distributed SystemsMicroservices at ScaleAdvanced Observability Vector Search TuningLLM Evaluation

Portfolio

Featured Projects

Commercial products first — a multi-tenant SaaS business platform, multi-branch POS systems and an LMS sold on CodeCanyon and Codester — followed by live client platforms: a B2B marketplace, an industry portal and high-traffic news systems. These are the Laravel, API, architecture and delivery work; the AI Engineering section covers the LLM, RAG and automation capability I bring to new builds.

SaaS · CodeCanyon

PayLink

Multi-tenant Laravel SaaS for agencies and service businesses — invoicing, recurring billing, projects, assets, documents and finance, sold on subscription.

Build

Tenant-scoped data layer behind super-admin, company and client portals. Queued jobs and the scheduler drive recurring invoices and reminders; Stripe, PayPal and bank transfer handle billing.

Role

Sole architect & lead engineer — tenancy and subscription model, every business module, gateway integrations.

Result

Published and actively maintained on CodeCanyon; buyers run it as their own hosted business platform.

POS · CodeCanyon

MB POS

Commercial multi-branch Point of Sale — inventory, sales processing, role-based access and reporting for retail businesses.

Build

Laravel REST API with a Vue.js register SPA so the counter stays responsive. Per-branch stock moves through an append-only ledger, keeping transfers and concurrent sales consistent.

Role

Sole architect & lead engineer — inventory engine, RBAC, installer, documentation and release.

Result

Published on CodeCanyon and deployed by customers on their own infrastructure.

LMS · CodeCanyon

LaraLMS

Learning management system with course authoring, enrolment, payments, quizzes, progress tracking and certificate generation.

Build

Modular Laravel platform separating courses, payments, assessment and reporting, with admin/instructor/student roles and a quiz engine tracking attempts and grading rules.

Role

Sole architect & lead engineer — course engine, quiz module and certificate generator.

Result

Commercial product on CodeCanyon used by training providers to run and monetise courses end to end.

POS · Codester

DevinTime POS

Multi-store Point of Sale for retail chains — one shared catalogue, per-store stock and staff, consolidated reporting.

Build

Laravel and Vue.js over a REST API, with store scoping enforced at the query layer so no endpoint or report can cross a store boundary.

Role

Sole architect & lead engineer — designed and built the system end to end, then published it commercially.

Result

Second POS product shipped to a developer marketplace, deployable by customers without engineering support.

B2B Marketplace

B2BMap

Live global B2B marketplace connecting suppliers and buyers through product listings, supplier discovery, inquiries and business messaging.

Build

Laravel modular monolith with domain boundaries for catalogue, suppliers, inquiries and messaging. Indexing strategy, query restructuring and Redis caching keep discovery fast as listings grow.

Role

Lead engineer — marketplace core architecture and the supplier–buyer communication flows.

Result

Running in production as a live marketplace with real businesses transacting through it.

B2B Portal

Textile Pages

Vertical B2B portal for the textile industry — manufacturers, suppliers and buyers with verified profiles and inquiry management.

Build

Domain-specific catalogue schema using indexed attribute tables rather than a generic EAV structure, with role separation across suppliers, buyers and moderators.

Role

Lead engineer & project planner — roadmap, data model and team delivery through to launch.

Result

Live industry portal that supports new categories without schema rewrites.

News Platform

Bangladesh24

National news portal covering breaking news, politics, business, sports and entertainment.

Build

Custom Laravel CMS with a draft → review → publish workflow and a cacheable read path invalidated on publish. Queued jobs handle media processing and distribution.

Role

Lead engineer — CMS, publishing pipeline and high-traffic content delivery.

Result

Stable publishing through traffic spikes, with editors shipping breaking stories unaided.

News Platform

EditorsTime

Editorial platform publishing current affairs, lifestyle, technology and cultural stories.

Build

Normalised tagging and category system with related-content resolution and indexed archive queries, so deep pages load as fast as the homepage.

Role

Lead engineer & project planner — platform architecture and content taxonomy.

Result

Archive content stays discoverable and SEO-visible long after publication day.

News Platform

NextNews24

24/7 news portal delivering continuous national, international, sports, entertainment and technology updates.

Build

Breaking-news pipeline with publish-triggered cache invalidation, queue workers for distribution and media, and scheduled health checks for continuous availability.

Role

Lead engineer — real-time update features and uptime planning.

Result

Continuous publishing with near-immediate content visibility and no scheduled downtime.

Production Backends, SaaS Platforms & AI-Powered Features

Tools & Technologies

Building AI Into Real Applications

Services

Engineering Capabilities by Domain

How I Build Systems

Technology Proficiency Matrix

Featured Projects

Academic Background

Let's Work Together