AI agents & automation engineer
Messy operations in.Autonomous systems out.
I build AI agents and automation systems that take the busywork off your team: lead qualification, data pipelines, CRM sync, and Q&A over your own documents. Shipped to production, not just demoed.
- 346K+
- verified records in production
- 300+
- users on a portal I built
- 50+
- workflows shipped
Live in production, not a demo.
Real operations problems, solved end to end.
Built a conversational AI agent that qualifies inbound student leads over chat (nationality, qualifications, program level, budget, and timeline) through natural back-and-forth rather than a rigid form. Structured field extraction runs in parallel, auto-classifies lead heat, logs everything to a live sheet, and hands hot leads straight to a human via instant WhatsApp alert.
Built a Google Apps Script automation that eliminated a manual nightly process entirely: candidate routing between a Master System sheet and multiple client sheets, running unattended on schedule. Zero human steps remaining.
Designed and built a hierarchical role-based portal from scratch, serving 300+ users across a five-tier organizational structure. Custom permission system, dynamic dashboards, and multi-sheet data orchestration, all in Google Apps Script.
Built an event-driven orchestration layer across Shopify, Stripe, and HubSpot. Handles payment verification, automatic retry on failed payments, CRM upsert logic, and warehouse Slack alerts. No human touchpoints in the primary path, from paid order to warehouse notification.
Agents and automations I've designed and shipped.
Ask the catalogue anything, get grounded answers.
Retrieval-augmented chatbot over a client equipment catalogue. Supabase pgvector for vector storage, Ollama nomic-embed-text for embeddings, and Groq for inference. Answers come from the catalogue itself, not from the model guessing. Deployed and verified in production.
Inbound leads enriched and briefed automatically.
A form submission triggers contact enrichment via Hunter.io, a recent-news lookup on the company, and an LLM-generated outreach angle tailored to that news. Everything lands in a CRM record and an internal email brief, ready to send with no manual research.
Raw listings in, categorized market signal out.
Scheduled pipeline pulls fresh listings, deduplicates by company and title, then runs each batch through an LLM classifier that tags category, top skills, and spam or anomaly flags. Results roll up into a weekly digest email with category breakdowns.
7-scenario Make.com orchestration for a clinic.
Seven coordinated Make.com scenarios handling appointment scheduling, patient follow-ups, staff notifications, and reporting. Replaced a fully manual operations workflow.
New client intake without the back-and-forth.
One form submission runs the full onboarding sequence: CRM record creation, an AI-drafted personalized welcome email, and a kickoff call scheduled on the calendar with the client already invited. No manual coordination.
Agents engineer.
Async-first.
Cairo.
I'm Mohamed, an AI agents and automation engineer. I design agents and workflows that remove manual work from real operations, using n8n, Make.com, and Zapier where they fit and custom Python where the no-code tools run out of road. My background spans BPO operations at Concentrix (Microsoft Azure) and Transcom (TUI UK), a degree in Management Information Systems, and several years of independent client work across Upwork and direct contracts.
I prefer text-based async communication. I work across time zones without friction. I don't bill by the hour for things that should run automatically.
Conversational agents that qualify leads, extract structured data, and hand off to a human at the right moment. RAG pipelines over your own documents for grounded Q&A. Gemini, Groq, and OpenAI as the reasoning layer.
Replacing manual processes with trigger-based workflows across your apps. n8n, Make.com, and Zapier as the execution layer, or custom Python where no-code hits its ceiling. If a human does it on a schedule, it can probably be automated.
Scrape, clean, enrich, push. Verified records at scale, running unattended on GitHub Actions or cron, landing in Supabase where your team can actually use them.
Let's work together.
Taking on select AI agent and automation projects. Email me the process that eats your team's time: what comes in, what has to happen, where it should end up. I reply within 24 hours. Async-first, calls on request.