Open for projects & remote roles

Jabed Hossain ·

Push to live. Sleep at night.

I set up deploy pipelines, lock down servers and keep watch on the alerts. Your team pushes code, it goes live, and nobody has to hold their breath. Dozens of sites run on setups I built. I also write the apps when needed, in Laravel, React, Next.js and TypeScript.

Start a project →Hiring? Grab my CV ↓

Press ⌘K or / to search · 1–7 to jump · ` for the terminal

jabed@byte: systemctl status

● ship.serviceactive (running)GitHub Actions · Linux ● guard.serviceactive (running)SSH · UFW ● operate.serviceactive (running)Bash · cron ● build.serviceactive (running)Laravel · Next.js

session uptime00:00:00

sites under my care29

domains watched13 · SSL ✓

01 / pipeline

The deploy flow behind every project I run. Press ▶ run to watch it ship, or click any stage to see what it does.

auto-deploy.yml · every repoready

stage/push

push · git

$ git push origin live
→ workflow: Auto Deployment
branches: stage → live (main on smaller repos)

# press ▶ run to deploy live

And when a release is bad?

Every deploy builds in its own folder and goes live by flipping one symlink. If the health check fails, the link flips back. Try the broken release.

release.sh · what happens on the server

ready

stagingdevelop branch
approvemerge + sign-off
backupdb dump · keep 5
buildnew release folder
switchcurrent → new (atomic)
healthcurl -f /health ≤ 10s
liverolling to every node

web-1:/var/www/app

currentreleases/20261011-1822

releases/

20261011-1822live
20261010-1105kept
20261009-1640kept
20261008-0912kept
20261007-1530kept

shared/ .env · storage/ · uploads/

Shared files survive every deploy. Each release is linked to them, never copied.

nginx · load balancer

web-11011-1822200 OK

web-21011-1822200 OK

web-31011-1822200 OK

One node at a time. A bad release stops at the first server, so the rest keep serving the old one.

# pick a release, then press ▶ deploy

Deploy a healthy release and watch it roll across the nodes.

Note: rollback switches code, not the database. Migrations stay backward compatible, and every deploy starts with a fresh dump.

02 / architecture

Not a one-off. It’s how PIF World, NLinBD, Shokkhom, Suniti, UIU and the rest all ship.

running in production

Every Laravel, React and Next.js project I work on ships the same way: push a branch, GitHub Actions takes over, deploys over SSH with a deploy script, and reports to Discord. I built this template and run it across RedOrange's projects.

19+repos on my CI/CD

8products shipping this way

2environments: stage · live

projects

PIF WorldNLinBDShokkhomSunitiUIU × IDSSShare-Net BDinternal tools

ci/cd

GitHub Actionstests · PR checks · deploy

SSH deploykey auth · deploy script

environments

live

stage

Discorddeploys · PRs

Sentryerrors (PIF)

what the template does

Branch-based auto-deploys: push to stage / live → GitHub Actions → SSH → deploy script
Secrets live in GitHub, servers accept key auth only
PR bot on every repo: merge-conflict checks + Discord alerts; unmerged PRs flagged automatically
Every deploy reports status, author, commit and run link to Discord
PHPUnit in CI where a repo has tests (SQLite, throwaway DB)
Docker Compose dev stacks (PHP-FPM, MySQL with health checks, Nginx)

biggest project on it

PIF World · the big one · playitforward.nl

My biggest project: a Dutch social-impact platform where nonprofits, companies and individuals run fundraisers, events, projects and petitions, and take donations through Mollie (iDEAL). I'm the main developer across its three codebases.

3,000+of my commits

3repos: api · front · admin

api: Laravel 12 · PHP 8.3 · MySQL. Sanctum auth, Google login, Mollie payments, roles & permissions, translations, PDF receipts
front: React · Vite · TanStack Query. i18n, charts, Google OAuth, Sentry
admin: React · Vite · Tailwind. content, donations and reporting for the team

Visit playitforward.nl ↗

03 / services

Pick one or run all four. Click a unit for its systemctl status.

Ship

From git push to production. No FTP, no prayers, no “works on my machine”.

● ship.service - Ship
   Loaded: loaded (/etc/systemd/system/ship.service; enabled)
   Active: active (running)
  ExecStart=GitHub Actions pipelines: build, test, deploy
  ExecStart=Linux server setup on AWS, Hetzner or cPanel/WHM
  ExecStart=Nginx, PM2, queues, cron, SSL
  ExecStart=Repeatable, scripted deploys and checklists

Good for Teams that still deploy by hand and hope

GitHub ActionsLinuxNginxAWS

Guard

Got hacked? Worried you will? I clean it up, then lock the doors.

● guard.service - Guard
   Loaded: loaded (/etc/systemd/system/guard.service; enabled)
   Active: active (running)
  ExecStart=Malware, backdoor & rootkit removal
  ExecStart=SSH hardening, key audits, UFW firewall
  ExecStart=Laravel & CMS hardening, Nginx rate-limits
  ExecStart=Package integrity checks and login audits

Good for Compromised servers, pre-launch security reviews

SSHUFWNginxdpkg

Operate

Backups, watchdogs and alerts, so the site stays up while you sleep.

● operate.service - Operate
   Loaded: loaded (/etc/systemd/system/operate.service; enabled)
   Active: active (running)
  ExecStart=Automated off-site backups (rclone → Google Drive) with retention
  ExecStart=Watchdog every 5 min, uptime + SSL/domain expiry checks
  ExecStart=Discord / email alerts with logs attached
  ExecStart=Patching, updates and monthly health reports

Good for Agencies and orgs running many sites on one box

BashcronrcloneDiscord

Build

I write the app too. Laravel on the back end, React and Next.js on the front, TypeScript throughout.

● build.service - Build
   Loaded: loaded (/etc/systemd/system/build.service; enabled)
   Active: active (running)
  ExecStart=Laravel APIs, auth, payments, admin panels
  ExecStart=React & Next.js front-ends in TypeScript / JavaScript
  ExecStart=Tailwind, TanStack Query, Vite, i18n
  ExecStart=MySQL, Oracle, MongoDB data models

Good for When you want one person from code to server

LaravelNext.jsReactTypeScript

04 / status

29 sites under my care 19 repos on my CI/CD

apps/Laravel / React apps on my CI/CD7

share-net/International knowledge network, one site per country8

gov & health/Government and public-health platforms4

programmes/NGO, rights and climate programmes7

redorange/Agency sites3

05 / incidents

Real incidents, written like postmortems. Clients anonymised, lessons kept. Click one to open it.

A production box hosting client sites had been quietly taken over. Even the ls command was in on it.

detectUnrecognised root logins and an SSH key nobody had issued.
investigate/usr/bin/ls replaced by a rootkit binary; a malicious cron file kept the door open.
eradicateRemoved backdoors, the rogue key and cron; reinstalled coreutils; verified packages with dpkg.
hardenPassword auth off, SSH hardened, UFW on, authorised IPs confirmed, audits automated.

root causeWeak SSH posture let an attacker in and persist.

lessonAssume the binaries can lie. Verify packages, not just processes.

$ ls -l /opt/opswhat runs on servers I look after

-rwxr-x---backup.shDatabases + files zipped, synced to Google Drive with rclone, 15-day retention
-rwxr-x---security.shPackage integrity (dpkg), SSH key and login audits, suspicious-account checks
-rwxr-x---server_maintenance.shUpdates, cleanup and service health
-rwxr-x---server_watchdog.shRuns every 5 minutes, restarts what falls over and alerts if it can't
-rwxr-x---domain_health.shUptime + SSL expiry across 13 domains
-rwxr-x---notifyDiscord and email alerts with attachments

06 / log

2024-10 → now

Full Stack Web Developer @ RedOrange Technologies

Runs and secures the company's multi-site Linux servers (Laravel, Next.js, React and CMS sites)
Automated backups, watchdog, security audits and alerting
Main developer on PIF World (playitforward.nl): API, front-end, admin and the CI/CD behind them
Built NLinBD, BMMDP, Suniti, Shokkhom

2024-05 → 2024-09

PHP & Laravel Developer @ IT Bangla Limited

PGCL (government), Notre Dame College applications
Pharmacy (BPDB) & school management systems

2023-06 → 2024-04

Junior PHP & Laravel Developer @ Leotech

Coastal & Marine Fisheries (Dept. of Fisheries), Hydrology (BWDB)
Schobee EMS; hospital software maintenance

2025 → now

MSc Software Engineering @ Northeastern University

Shenyang, China

2019 → 2023

BSc Software Engineering @ Daffodil International University

Dhaka

$ apt list --installed

ops/

linuxnginxdockergithub-actionsci/cdsentrypm2cronrcloneufwssh

cloud/

awshetznercpanel/whm

languages/

typescriptjavascriptphpbashsql

frontend/

reactnext.jstailwindtanstack-queryvitei18next

backend/

laravelsanctumrest apisqueues

data/

mysqloraclemongodbsqlite

certs/

i2i Virtual Internship: PHP with Laravel
Certificate in Software Development (PHP with Laravel)

Push to live. Sleep at night.

push · git

And when a release is bad?

PIF World · the big one · playitforward.nl

Ship

Guard

Operate

Build

Full Stack Web Developer @ RedOrange Technologies

PHP & Laravel Developer @ IT Bangla Limited

Junior PHP & Laravel Developer @ Leotech

MSc Software Engineering @ Northeastern University

BSc Software Engineering @ Daffodil International University

Need it built, shipped or rescued?

Hiring a DevOps engineer?

Push to live. Sleep at night.

From git push to green.

push · git

And when a release is bad?

Same pipeline. Every repo.

PIF World · the big one · playitforward.nl

Four services. One human.

Ship

Guard

Operate

Build

Uptime is a feature.

Things broke. Here’s how I fixed them.

git log --author=jabed

Full Stack Web Developer @ RedOrange Technologies

PHP & Laravel Developer @ IT Bangla Limited

Junior PHP & Laravel Developer @ Leotech

MSc Software Engineering @ Northeastern University

BSc Software Engineering @ Daffodil International University

Let’s ship something.

Need it built, shipped or rescued?

Hiring a DevOps engineer?