Skip to content
Menu

About

I build the system, not just the prompt.

I'm Mahmoud Samir Atieh, an AI Automation Engineer based in Istanbul with 1+ year of experience in AI automation. I build retrieval systems, AI agents and automated workflows — and the backend, data model and deployment that let them run in front of real users.

My work tends to start with a process someone is doing manually: answering the same customer questions, reading the same documents, moving data between tools. I map what that process actually is, then build the smallest reliable system that takes it over — with the failure modes handled, because a system that silently guesses is worse than no system.

I care about grounding. In every project here, an answer can be traced back to something real: a page in an uploaded document, a product record, an order, a log. That constraint shapes the architecture more than any model choice does.

Email
mahmoudsamiratieh@gmail.com
Focus
AI agents · Retrieval · Automation
Working
Remotely, with clients abroad

Experience

  1. AI Specialist & IT Support

    Jun 2025 — Jan 2026

    Media Vision · Internship

    • Built a news chatbot on a custom RAG pipeline so answers stayed grounded in real reporting rather than model memory.
    • Designed and tuned system prompts for tone, factual accuracy and structured output inside automated news workflows.
    • Built an image-generation workflow that produces branded visuals from company templates automatically.
    • Maintained internal AI tooling and supported the team's day-to-day technical needs.
  2. B.Sc. Software Engineering

    2026

    Istanbul Sabahattin Zaim University

    • Graduation project: LYRA, a smart-home voice assistant that runs entirely offline — local speech recognition, a layered NLP cascade with a local model, and local speech synthesis.
    • Autonomous robotics software in Python and ROS2 for Turkey's national Teknofest competition.

Engineering Arsenal & Stack

AI & RAG Architecture

Core Engine

1536-dim Vector + BM25

Production LLM orchestration, hybrid dense/sparse vector retrieval, semantic chunking & grounded citation auditing.

OpenAI GPT-4oPrimary LLM
Google Gemini 2.0Multimodal
Custom RAG PipelinesArchitecture
Embeddings & Vectorspgvector
Hybrid SearchDense + Sparse
Cross-Encoder Re-rankingPrecision
Document OCRVision & Layout
Structured OutputsZod / JSON Schema
Zero Hallucination Grounding Standard

Agentic Automation

Orchestration

LAYER 02

Autonomous tool-calling loops, multi-step webhook triggers, event streaming and business integrations.

n8n Self-HostedREST APIs & WebhooksWhatsApp Business APIShopify Admin APIOAuth 2.0 Auth FlowTelegram Bot API
Integration: n8n + Custom Webhook Nodes

Backend & Data

System Runtime

LAYER 03

Type-safe APIs, relational data modeling, migrations, background workers and edge datastores.

Node.js & TypeScriptPython 3.11+PostgreSQL & pgvectorPrisma ORMSupabaseREST & GraphQL
Stack: PostgreSQL · Prisma · Supabase · Python

Cloud & DevOps

Infrastructure

LAYER 04

Isolated container runtimes, automated CI/CD pipelines, DNS proxying, SSL and high-availability VPS nodes.

Docker ContainersLinux (Ubuntu Server)Cloudflare Workers & DNSVercel Edge NetworkHostinger VPSGit & GitHub Actions
Infra: Docker · Linux VPS · Cloudflare Workers

Spoken Languages

Arabic (Native) · English (Advanced) · Turkish (Intermediate)