Abdul Rehman Zahid
I build AI agents, RAG systems, and multilingual NLP that survive production
Three years of taking language models the last mile: agents with memory that actually hold a conversation, retrieval over documents nobody wanted to read, and the boring infrastructure underneath that keeps both of them awake at 3am. Currently at Xeven Solutions. On the side I teach AI engineering one-to-one on Preply — 50+ students, 400 lessons, project-based from LLM APIs through RAG, agentic design, and n8n automation.
I design and deploy production AI agents and multi-agent systems with long-term memory across legal tech, education, and enterprise workflows — cutting manual operational effort by 40%. I architected the RAG stack (semantic search, document processing, vector storage) on OpenAI, Claude, and Gemini, and the orchestration layer in LangChain and LangGraph with MCP, Slack, Gmail, and LinkedIn integrations. I also own how it runs: Docker, AWS EC2, self-hosted n8n behind Nginx and PM2, redundant vector databases, API failover, and LangSmith monitoring for observability and spend.
50+ students worldwide, 400+ lessons — project-based teaching on LLM APIs, prompt engineering, Claude Code, RAG, agentic AI design, and n8n automation.
Structured large volumes of operational and customer data, built reusable templates and database workflows that cut manual errors, and standardized reporting formats across departments. This is where the automation instinct started.
Government College University Faisalabad. Engineering fundamentals first; the machine learning came from evenings, papers, and building things that didn't work until they did.
Abdul Rehman Zahid is an AI engineer based in Faisalabad, Pakistan. At Xeven Solutions he designs production-grade AI agents and multi-agent systems with long-term memory, and architected the company's retrieval stack — semantic search, document processing, vector storage — across the OpenAI, Anthropic Claude, and Google Gemini APIs. His Arabic NLP platform has processed over a million multilingual documents at 92% accuracy.
He also upgraded the company's deployment practice — Docker, AWS EC2, redundant vector databases, API failover, LangSmith observability — and he teaches AI engineering to students worldwide on Preply. He holds a BSc in Electrical Engineering Technology from Government College University Faisalabad.
- Code and experiments on GitHub.
- Work history and posts on LinkedIn.
- Reach me at arzaiengineer@gmail.com — open to AI engineering roles and consulting.
- English and Urdu; comfortable building for Arabic-language products.