a
azamali992

Azam Ali Afzal

@azamali992

AI Agents and Business Automation Engineer

Pakistan
Anglais, Ourdou
Certaines informations sont présentées en anglais.
À propos de moi
I help businesses reduce manual work with AI agents, document processing and workflow automation. I build systems that turn messy inputs into structured data, connect tools through APIs and keep people in control of important actions. My work covers Python, FastAPI, LangChain, LangGraph, OCR, RAG and n8n. I include testing, clear setup instructions and honest limitations so you can understand and maintain what I deliver. I am an MS AI candidate at LUMS and co-founder of AZEKTRA. Share your workflow and the task you want to automate, and I will help define a practical scope.... Plus d’infos

Compétences

a
azamali992
Azam Ali Afzal
hors ligne • 
Temps de réponse moyen de 1 heure

Voir mes services

Automatisations
I will automate sales reports with source checks and exception exports
De100 $US / projetPlus de détails
Sites web IA & Logiciel
I will set up an ai agent based erp for your distribution business
De200 $US / projetPlus de détails

Portfolio

Expérience professionnelle

Multan_Chemicals

ai consultant

Multan Chemicals • Freelance

Jun 2026 - Sep 2026 • 3 mos

Since June 2026 I've been working as an AI and automation consultant for Multan Chemicals Limited (MCL), a chemicals manufacturer in Pakistan that delivers its products to dealers and clients in cylinders. MCL runs its business on Oracle (APEX, cloud hosted), but a lot of the actual work around it was still happening on paper. Orders were written down, bills were handwritten, drivers carried paper slips to clients, and at the end of it a data entry team sat and typed everything back into Oracle by hand. My job was to find where that paper was costing them time and accuracy, and replace it with software that talks to the system they already have. I designed and built two systems there: a full digital delivery pipeline with role-based mobile apps integrated with Oracle, and a document pipeline that reads their scanned ledger pages, works out what each page is, and files it under a proper name. The delivery system was the bigger of the two in terms of how many people it touches. Before I started, the flow looked like this. Production wrote a bill identified by an ECR number. The driver took that ECR along with the cylinders, the client signed a paper delivery challan, and the paper came back to the office where someone re-keyed it into Oracle. Every handoff was a place where data could go wrong. A number copied incorrectly, a quantity missed, a slip lost in the vehicle, a cash amount that didn't match what was written. And because the data entry happened days later, nobody had a live picture of what had left the warehouse, what was on the road, and what had been delivered. On top of that, the ECR numbers themselves couldn't be trusted as keys. MCL has several warehouses and around five different product book types, and each book restarted its numbering at 1 every financial year. So the same ECR number existed many times over, in different warehouses and different books. You can't build a reliable digital system on an identifier that repeats. One of the first things I