AI Engineer · Linux Guy (DevOps) · Software Engineer
I design and ship end-to-end intelligent systems, from Arabic-aware NLP pipelines to multi-agent orchestration, using FastAPI, LangGraph, and Google Gemini. Currently building production AI platforms for Jordanian government institutions at 9XAI / HTU.
I'm an AI Engineer based in Zarqa, Jordan, specializing in Retrieval-Augmented Generation (RAG) systems, multi-agent AI architectures, and LLM-powered application development. My work spans prompt engineering, AI agent orchestration, and building end-to-end intelligent systems with FastAPI, LangGraph, and Google Gemini. I apply vibe-coding and AI-assisted development to rapidly deliver full-stack platforms for government and enterprise clients, turning complex AI capabilities into usable, production-grade products. I'm a fellow in the 9XAI program at Al Hussein Technical University (HTU), and the creator of 9XAIPal, my personal AI research and reading companion. Our leader, Rami Alkarmi, is the founder of 9XAI.
Python, SQL, Bash, JavaScript
PyTorch, Transformers, Scikit-learn, AI Agents, RAG Systems, LangGraph, LangChain, ChromaDB
Google Gemini (Vertex AI), Qwen, LLaVA, Ollama, Prompt Design, Vibe-Coding
ETL Pipelines, Apache Kafka, Web Scraping, SQLAlchemy, MySQL
FastAPI, Flask
HTML, CSS, Streamlit (AI-assisted full-stack via vibe-coding)
PostgreSQL, MySQL, MongoDB, SQLite, Neo4j / NebulaGraph, Redis
Docker, Git/GitHub, AWS, Linux, Shell Scripting, Vercel
A few of these were built for government institutions under NDA, so descriptions stay at a high level and source code stays private. See public work on GitHub
A regulatory intelligence platform that helps legal and policy teams keep up with fast-changing Arabic and English legislation. It automatically pulls out obligations from new laws, flags where regulations conflict with one another, and tracks how legislation changes over time, cutting down work that would otherwise mean manually cross-referencing hundreds of documents.
A unified oversight platform that gives government decision-makers one place to see how public services are performing across 14+ Jordanian ministries, covering education, health, finance, energy, justice, and the digital economy. It reads through ministry documents with RAG-based search so leaders can ask direct questions instead of digging through reports.
My personal AI research and reading companion for students and researchers working with Arabic and English papers, technical books, and articles. It turns dense documents into readable articles, lets you ask about passages, figures, and equations with source citations, and keeps notes beside the material they refer to. Available as a free hosted beta and an open-source, self-hosted app with support for local or cloud models.
Visit 9XAIPal About the research companionA satellite-data feasibility study for Jordan's Gulf of Aqaba marine reserve, built to inform ASEZA's coastal management and monitoring decisions. Evaluated 14 satellite and ocean data sources across a 24-year chlorophyll record and a 42-year thermal archive to establish what can, and cannot, be reliably measured from space, then shipped a validated 7-day chlorophyll forecast that consistently beats the seasonal baseline. Built end-to-end by a 5-person team in a 9-day sprint.
View on GitHubJordan's national Citizen-Experience Intelligence platform. It ingests citizen feedback from every government channel, anonymizes personal data before it enters the pipeline, and runs it through an Arabic-native NLP/ML pipeline to produce a single explainable Citizen Experience Index, ranked root-cause recommendations with named owners, and a public portal where citizens can submit, track, and ask about their cases. Built sovereign by design, with all processing kept inside the government boundary.
A conversational assistant that lets people ask natural-language questions, in Arabic or English, about Jordan's national vision documents and get answers grounded in the actual source text. Low-confidence answers are routed to a human-in-the-loop review step instead of being guessed, so accuracy holds up even on ambiguous questions.
B.S. in Data Science and Artificial Intelligence
Completed the degree in 3.4 years, ahead of the standard track
Oct 2022 - Jan 2026 · Zarqa, Jordan"How I Turned My Weaknesses into Strength"
HTU, Al-Hussein Technical University · 2026Khaled Saleh is an AI engineer based in Zarqa, Jordan, a fellow in the 9XAI program at Al Hussein Technical University (HTU), and the creator of 9XAIPal. He specializes in Retrieval-Augmented Generation (RAG), multi-agent AI systems, and Arabic-aware applications.
Founded and led by Rami Alkarmi, 9XAI is an applied AI program at Al Hussein Technical University (HTU) in Jordan. It connects young technical talent with government institutions to identify real public-sector challenges and develop AI-powered solutions, supporting public-sector modernization.
Read about 9XAI at HTU →Rami Alkarmi (also written Rami Al Karmi) is the founder and leader of the 9XAI program and our fellowship team. He serves as President’s Assistant for Innovation at Al Hussein Technical University (HTU). An innovation strategist, his background spans AI, entrepreneurship, fintech and digital learning.
Rami’s role at HTU → Read Rami’s biography →9XAIPal is an AI research and reading companion created by Khaled Saleh for students and researchers. It helps you read research papers, technical books and articles in Arabic and English, ask questions with citations to source passages, organize notes and search across your library. It is Khaled's personal project; Khaled is also a 9XAI fellow at HTU.
Learn about 9XAIPal →I'm open to opportunities in AI engineering, RAG systems, and multi-agent architectures. Reach out directly or connect below.