AI Automation Engineer

I build AI systems and products that ship — not demos.

I design and deploy production software: multi-agent LLM pipelines grounded in real data, full-stack platforms with real payments and video, and the automation that keeps them reliable. Three projects below are live right now — click through and use them.

View Projects Download Resume

Rawalpindi, Pakistan azam.lawa@gmail.com github.com/MAzam1999

Core Stack

TypeScript Next.js 16 / React 19 PostgreSQL + pgvector Prisma / Drizzle ORM Multi-agent LLM pipelines Stripe Clerk / Supabase Auth Mux Video Docker AWS EC2 / ECR GitHub Actions CI/CD n8n Automation

Selected Work

Three flagship projects, not eleven shallow ones

Every project below is a real, deployed product. Depth over breadth — see the full build story on each case study.

QuranicLens

Multi-agent RAG · Islamic knowledge base

Live

Explores any Quranic verse through six lenses — tafsir, hadith, Seerat, science, history, and scripture parallels — synthesized by a 4-agent LLM pipeline grounded in a 78,000+ record verified knowledge base so it never invents a source.

Next.js 16 pgvector 4-agent pipeline SSE streaming

Case Study Live Demo

FreshTrack

Restaurant inventory · Mobile-first PWA

Live

Batch-level FEFO inventory tracking for restaurants — barcode scanning, expiry alerts, waste analytics, and rule-based reorder forecasting. Built as an installable PWA, mobile-first end to end.

Next.js 16 Drizzle Barcode scan PWA

Case Study Live Demo

LMS Platform

Course platform · Payments + Video + CI/CD

Live

A full course-selling platform — instructors publish video courses, students pay and track progress. Ships to production two ways from one codebase: Vercel, and a self-hosted Docker container on AWS EC2 via a GitHub Actions pipeline.

Prisma + Neon Stripe Mux Video Docker / EC2

Case Study Live Demo

Experience

Where the work has been

AI Automation Engineer — Independent

2024 — Present · Remote

Designed and shipped QuranicLens, FreshTrack, and the LMS Platform end to end: architecture, data model, AI pipeline, and deployment
Built multi-agent LLM pipelines with grounded retrieval (pgvector) instead of relying on model training data
Shipped production payments, video, and role-based access control (Stripe, Mux, Clerk)
Set up CI/CD pipelines shipping the same codebase to both Vercel and self-hosted Docker/AWS

Data Science Intern

The Spark Foundation · 09/2023 – 10/2023 · Remote

Designed and optimized machine learning models to improve prediction accuracy
Built data visualizations to turn complex datasets into actionable insight

Data Science Intern

Omdena · 08/2023 – 09/2023 · Remote

Cleaned and preprocessed large datasets for a fake-news detection model
Ran exploratory data analysis for a weather-forecasting initiative with a global team

Workflow Automation Specialist — Freelance

12/2019 – 01/2023 · Remote

Designed end-to-end workflow automation, reducing manual effort across client processes
Integrated AI tools into existing systems for data collection, analysis, and reporting

Get in touch

Let's build something that ships

Open to freelance and full-time AI automation work.

azam.lawa@gmail.comEmail +92 301 8591856Phone Muhammad AzamLinkedIn MAzam1999GitHub