Skip to content
Open to full-time roles and freelance projects

Farooq Shabbir / Full Stack DevOps Engineer · AI First I ship the platform, the pipeline and the AI feature.

Full Stack DevOps Engineer working AI-first. I build backends in Python and FastAPI, frontends in Next.js, and the AWS and Azure infrastructure, CI/CD and release tooling that get them to production safely.

Start a projectSee selected workLahore, Pakistan · remote
Clouds: AWS, Azure, GCP
03
Clouds: AWS, Azure, GCP
Job platforms automated by the agent
04
Job platforms automated by the agent
Dimensions in the LLM-as-judge rubric
05
Dimensions in the LLM-as-judge rubric
Downtime for production index changes
00
Downtime for production index changes

Pipeline · scroll to deploy

Every commit takes the same road to production.

Scroll to ship one. Fast feedback first, signed artifacts, and a deploy that has to prove itself before it counts.

elapsed

T+00:00

waiting for push
  1. commit
  2. build
  3. test
  4. sign
  5. stage
  6. deploy
  7. verify

Selected work

Systems I've designed, shipped and kept running.

Production work from the last year, from a desktop agent's release pipeline to real-time AI. Client names are kept private; the engineering is shown as it was built.

01 / 04 · Release engineering

A self-updating desktop agent that can't be tampered with

A Windows agent that installs its own updates has to prove every byte is genuine and recover when a release goes wrong.

  • Ed25519-signed update manifests and SHA-256 checks on every artifact
  • Automatic rollback when a new build fails to start
  • Packaging and startup reworked so Microsoft Defender stops quarantining it
  • OV code signing via cloud signing, PyInstaller and Inno Setup
agent · self-updatechecking

  1. fetch manifest.json

    over HTTPS

  2. verify Ed25519 signature

    public key pinned in agent

  3. download + SHA-256 check

    hash must match manifest

  4. stage and swap binary

    single-instance lock held

  5. health check new build

    start within timeout

✓ healthy → keep
↺ fails → auto-rollback

02 / 04 · Applied AI

A real-time voice interviewer, tuned for latency

Speech-to-text, an LLM and text-to-speech streamed over WebSockets, with an LLM-as-judge scoring each answer on a five-dimension rubric.

  • Parallel save path so persistence never blocks the reply
  • Redis pre-check before expensive work
  • Speculative follow-up audio and a pre-warmed speaker
  • Claude API for questions, scoring and cover letters
voice-interview · one turn
speech → text
save answer
redis pre-check
LLM first token
text → speech
reply audible

Illustrative: work that used to run in sequence now overlaps, so the reply starts sooner.

03 / 04 · Cloud infrastructure

AWS networks built to be rebuilt

VPCs, load balancing and auto scaling defined end to end and written up as runbooks, so any engineer can reproduce an environment.

  • Public and private subnets, NAT and Internet Gateways
  • ALB path routing, health checks, ACM TLS and HTTP→HTTPS
  • Auto Scaling Groups from custom AMIs and Launch Templates
  • IAM roles per instance, SSM instead of SSH
vpc · productionpublic + private tiers
users · HTTPSRoute 53DNS delegationACMTLS certificateVPCpublic subnetsALBpath routing · healthNAT GWIGWprivate subnets · auto scaling groupEC2custom AMIEC2custom AMIEC2scaling…IAM role per instance · SSM, no SSH · Secrets Manager

04 / 04 · CI/CD

Pipelines that fail loudly, not silently

GitHub Actions and Azure DevOps pipelines that take a commit all the way to a verified deployment, and stop the moment something is wrong.

  • Webhook and path-filtered triggers on self-hosted agents
  • Build, sanity checks, deploy and post-deploy verification
  • Polling that fails the run on error or timeout
  • Separate staging for client acceptance before production
pipeline · commit → productionself-hosted agent
1
trigger
2
build
3
sanity
4
deploy
5
verify

on: push · paths: ["api/**", "infra/**"]

verify: poll /healthz every 10s

fail_if: error || timeout 300s

… waiting on stages

Experience

Where I've done the work.

  1. Apr 2026 — Present

    Current

    Full Stack DevOps Engineer (AI First) · Freelance

    Remote · UK-based client

    Backend, frontend, Windows desktop agent and release infrastructure for an AI-powered job-application automation platform, built with an AI-first workflow (Claude Code) and strict review, testing and verification on every change.

    Show 8 highlights
    • Built and maintain a Windows desktop agent that fills applications across Greenhouse, Lever, Ashby and LinkedIn with Playwright, with a review step before submission and sensitive data redacted from logs.
    • Designed a secure self-updater: Ed25519-signed manifests, SHA-256 integrity checks, automatic rollback on a failed build and single-instance protection, validated end-to-end on the packaged build.
    • Resolved Microsoft Defender quarantining the agent by changing how it is packaged and launched at startup; own the code-signing strategy (OV certificate, cloud signing).
    • Sped up the API and database (FastAPI, PostgreSQL/Supabase) with targeted queries and zero-downtime indexes; added guaranteed agent-to-server result delivery and remote device revocation.
    • Shipped AI features: Claude-generated interview questions, an LLM-as-judge scoring answers on a five-dimension rubric, a streaming voice interview (STT → LLM → TTS over WebSockets) and generated cover letters.
    • Cut voice-response delay with a parallel save path, a Redis pre-check, speculative follow-up audio and a pre-warmed speaker.
    • Architecting the AWS deployment: GitHub Actions CI/CD, a separate staging environment for client acceptance, secrets in AWS Secrets Manager, HTTPS, uptime monitoring and recovery runbooks.
    • Configured transactional email (Resend with Supabase Auth) on an isolated subdomain to protect sender reputation, and work directly with the founder on requirements, cost and releases.
    • Python
    • FastAPI
    • Next.js
    • PostgreSQL
    • Supabase
    • Redis
    • Playwright
    • Claude API
    • AWS
    • GitHub Actions

  2. Jul 2026 — Aug 2026

    DevOps Engineer Intern · NetSol Technologies

    Lahore, Pakistan

    CI/CD, containerisation and AWS networking work, documented so another engineer can rebuild each environment from scratch.

    Show 6 highlights
    • Built CI/CD pipelines in GitHub Actions and Azure DevOps (YAML, self-hosted agents) triggered by webhooks and path-filtered commits, from checkout through build, sanity checks, deploy and post-deploy verification.
    • Replaced manual steps with one-command Bash automation for Docker build-and-push, service deployment and environment setup and teardown.
    • Packaged apps as multi-stage Docker images with Docker Compose and nginx, and ran them on Ubuntu with systemd.
    • Designed VPCs with public and private subnets, NAT and Internet Gateways, Route 53 delegation, ACM certificates and ALB path-based routing with health checks and HTTP→HTTPS redirects.
    • Provisioned and scaled EC2 with custom AMIs, Launch Templates and Auto Scaling Groups; monitored with CloudWatch alarms, SNS alerts and health endpoints.
    • Applied least-privilege IAM roles per instance, kept secrets in AWS Secrets Manager and Azure DevOps, and replaced SSH with SSM Session Manager.
    • GitHub Actions
    • Azure DevOps
    • Docker
    • nginx
    • AWS VPC
    • EC2
    • ALB
    • Route 53
    • CloudWatch
    • Bash

  3. Jan 2026 — Jun 2026

    DevOps Intern (CMIT) · Argon Teq Limited

    Lahore, Pakistan

    First production cloud work: deploying and monitoring a client recruitment application under senior engineers.

    Show 3 highlights
    • Deployed a Node.js and Next.js recruitment application for a client on AWS (EC2, VPC, Docker).
    • Supported GitHub Actions pipeline configuration and built and tested Docker images for microservices.
    • Monitored performance with AWS CloudWatch and Azure Monitor and flagged optimisation opportunities.
    • AWS EC2
    • VPC
    • Docker
    • GitHub Actions
    • CloudWatch
    • Azure Monitor

Freelance services

Bring me in for the part that has to work.

Fixed-scope engagements or ongoing support for startups and teams. Every engagement ends with documentation another engineer can follow.

01

CI/CD pipelines

Commit-to-production pipelines on GitHub Actions, Azure DevOps, GitLab or Jenkins, with staging, checks and post-deploy verification.

  • Pipeline as code
  • Staging environment
  • Runbook
02

Cloud infrastructure on AWS and Azure

Secure networks, load balancing, auto scaling, TLS and DNS, with least-privilege access and secrets kept out of the repo.

  • VPC and IAM design
  • HTTPS and DNS
  • Cost review
03

Containerise and deploy

Multi-stage Docker images, Compose, nginx and systemd on Linux, moving an app from a laptop to a reliable server.

  • Dockerfiles
  • One-command deploy
  • Health checks
04

AI features in production

LLM features built on the Claude API: generation, LLM-as-judge evaluation and streaming voice, designed for latency and cost.

  • Prompt and rubric design
  • Streaming backend
  • Evaluation
05

Desktop release and code signing

Packaged Windows apps with signed, verifiable auto-updates and rollback, and fixes for antivirus false positives.

  • Installer
  • Signed updates
  • Rollback
06

Monitoring and runbooks

CloudWatch and Azure Monitor alerting, health endpoints, and the recovery and backup runbooks your team will actually use.

  • Alarms and alerts
  • Recovery runbooks
  • Backup plan
1Discovery call2Written plan and quote3Build with weekly demos4Handover and runbooks

Stack

Tools I use in production.

Cloud
AWS · Azure · GCP
CI/CD
GitHub Actions · Azure DevOps · AWS CodePipeline · CodeBuild · CodeDeploy · Jenkins · GitLab
Languages & frameworks
Python · FastAPI · SQLAlchemy · Next.js · React · TypeScript
Data
PostgreSQL · Supabase · Redis
Containers & OS
Docker · Docker Compose · nginx · Linux · Windows
Automation
Playwright · Patchright · Chrome DevTools Protocol · Bash · PowerShell
Release
PyInstaller · Inno Setup · Code signing · Ed25519 updates · SHA-256 integrity
Security
AWS IAM · Secrets Manager · Security Groups · Azure RBAC · Defender compliance
Observability
CloudWatch · Azure Monitor · Log Analytics · Application Insights

About

Engineer first. AI-first by workflow.

I work across the whole path a product takes to reach users: the Python and FastAPI backend, the Next.js frontend, the cloud it runs on, the pipeline that ships it and the monitoring that tells you when it breaks.

I build with Claude Code as part of my daily workflow, and pair that speed with strict review, testing and verification on every change. I'm currently completing an MS in Computer Science at FAST alongside client work.

Education

  • MS Computer Science

    National University of Computing and Emerging Sciences (FAST)

    2025 — 2027In progress · CGPA 3.32

  • BS Computer Science

    University of Central Punjab

    2020 — 2024CGPA 3.34

Certifications

  • Oracle Multi Cloud Professional
  • Huawei HCCDP Solution Architect
  • DevOps and AI on AWS (Coursera)

Award

Winner, Code Rush 2024

University of Central Punjab coding competition, ahead of 100+ participants with a 95% completion rate.

Contact

Let's build something reliable.

Hiring for a DevOps or full-stack role, or need a pipeline, cloud setup or AI feature shipped? Send a note and I'll reply personally.

  • Open to full-time roles and freelance work
  • Remote from Lahore, Pakistan (UTC+5)
  • Your details are only used to reply to you
I'm reaching out about