Automation & agents
n8n workflows, Telegram and Slack bots, glue between SaaS tools, scheduled jobs that don't silently rot.
Independent engineer · MENA timezone · remote
Workflows that don't break at 2am. Data pipelines you can actually audit. LLM features that survive contact with real users. A decade-plus across enterprise stacks, now shipping for founders and small teams who need senior judgement without the headcount.
n8n workflows, Telegram and Slack bots, glue between SaaS tools, scheduled jobs that don't silently rot.
Production-grade Claude / Gemini / OpenAI integrations. Prompt evals, retries, caching, observability — not demos.
Schemas, ETL, dashboards, ERP and HR data work. The unglamorous middle layer that makes leadership actually see the business.
30-minute call. We map the actual workflow on a whiteboard before any code or proposal — most engagements get smaller, not bigger, after this step.
I ship the smallest useful slice first — usually inside a week — running on your infra. You judge it on real data, not a deck.
Documented, tested, deployed under accounts you own. No silent vendor lock-in. If you bring an in-house engineer in six months, they can read the code.
Three-phase Docker stack — research → script QA → image gen → TTS → Remotion render. n8n + Python + Postgres. Replaces a 4-person editorial team for one channel.
Cloudflare Worker pulls structured data from regional customs PDFs via Vertex AI. Built and deployed in a working week. ~86% per-field confidence on first batch.
53-slide medical seminar reset for AR/EN, brand-true typography, exported PDF + editable masters. Three-hour turnaround, bounded LLM cost.
References available on request — most live work is under NDA.
Best way to reach me is email — short note about what you're trying to build is enough. I usually reply within 24 hours, MENA business days.
Prefer something else? Add a contact handle, your timezone, and a one-line description of the work — I'll route it.