NextAura AI
Back to founders

Founder profile

Muayid Rabah

Founder & AI Engineer · LLMs, RAG, MCP Agents & Computer Vision · Python/FastAPI Developer

Amman, Jordan

AI engineer and recent AI graduate who builds practical products end to end across RAG, LLMs, computer vision, machine-learning APIs, Android, and full-stack AI. His work includes FitCoach AI, deployed ML APIs, computer vision, and training in Claude APIs, MCP, Amazon Bedrock, Google Cloud Vertex AI, and agent workflows.

Very Good
Education
5
Selected projects
7
Technical toolkit
18
Certifications & training
Muayid Rabah, Founder & AI Engineer · LLMs, RAG, MCP Agents & Computer Vision · Python/FastAPI Developer
Current role

Founder & AI Engineer · NextAura AI

July 2026 – Present • Amman · On-site / Hybrid

Muayid designs and implements intelligent product capabilities, including RAG systems, LLM integrations, AI agents, computer vision, ML APIs, and production-ready AI workflows.

  • Builds chatbots, RAG pipelines, and LLM-powered digital products.
  • Develops machine-learning services, computer-vision capabilities, Python/FastAPI backends, and full-stack AI products.
  • Supports AI consultation, automation, website MVPs, and product architecture.
  • Researches MCP, subagents, Amazon Bedrock, Google Cloud Vertex AI, and agent workflows.
Selected projects

FitCoach AI — Intelligent Virtual Personal Trainer

A bilingual AI fitness coach using RAG, LLMs, machine learning, fitness data, health conditions, meal analysis, workout guidance, and follow-up.

ReactFastAPIRAGFAISSLLMSentenceTransformers
  • Implemented FAISS and SentenceTransformers RAG, LLM integration, backend APIs, preprocessing, validation, and safety handling.
  • Contributed product architecture, health recommendation logic, progress tracking, admin follow-up, and bilingual Arabic/English support.
  • Built meal-photo and medical-report analysis, workout browsing and scheduling.

Family Vault — Android Password & Notes Manager

An offline Android app for household passwords, PINs, and private notes.

KotlinAndroidFragmentsViewPager2Navigation ComponentRecyclerView
  • Built Fragment navigation with Safe Args, a two-tab ViewPager2 experience, RecyclerView CRUD, and tap-to-reveal behavior.
  • Tested with three family members, with average usability above 4/5.

Cat vs Dog Classification — Hybrid Deep Learning

A computer-vision classifier trained on more than 10,000 images, combining deep learning with image-processing techniques.

PyTorchEfficientNetB0OpenCVPython
  • Fine-tuned EfficientNetB0 and used Harris corner detection, ORB, Haar cascades, template matching, histogram equalization, and Canny processing.
  • Built real-time annotated inference and image preprocessing.
93.5% accuracy0.9354 macro F12.3 percentage-point preprocessing improvement

House Price Prediction API

A FastAPI machine-learning API built during the Ministry training using California socioeconomic data, from cleaning and feature engineering through evaluation, packaging, and a live endpoint.

FastAPIscikit-learnPython
  • Prepared data, engineered features, evaluated the model, packaged it, and deployed a live prediction endpoint.

Student Academic Performance Analysis

An analysis of 1,044 UCI student records comparing Random Forest, SVM, Naive Bayes, Logistic Regression, and kNN to identify factors associated with academic performance.

PythonOrangeRandom Forestscikit-learn
  • Compared models and identified absences, failures, and age as relevant factors.
Random Forest: 79.5% accuracyAUC: 0.806
Experience

AI Trainee

Ministry of Digital Economy and Entrepreneurship, Jordan

September 2025 – November 2025 · Amman, Jordan · Hybrid · Internship

Applied AI training focused on a live house-price FastAPI endpoint and collaborative development practices.

  • Used raw socioeconomic data for cleaning, feature engineering, model evaluation, packaging, and deployment.
  • Practiced prompt and LLM concepts alongside Git branches, pull requests, and team collaboration.
Technical toolkit

Languages

PythonJavaScriptTypeScriptKotlinJavaSQL

AI Agents & LLMs

LLMsRAGFAISSSentenceTransformersMCPSubagentsPrompt EngineeringClaude APIAmazon BedrockGoogle Cloud Vertex AI

Machine Learning & Computer Vision

PyTorchscikit-learnEfficientNetOpenCVFeature EngineeringCross-validationAUCF1Image PreprocessingDeep LearningComputer Vision

Web & APIs

FastAPIReactRESTHTMLCSS

Mobile

AndroidFragmentsViewPagerNavigationRecyclerViewMaterial Design

Cloud & Tools

AWSGitGitHubAndroid StudioVS CodeJupyterPostmanOrange

Data

pandasNumPyMatplotlib
Education

Bachelor of Information Technology — Artificial Intelligence

Zarqa University, Jordan

February 2022 – June 2026 · Very Good

Machine LearningApplied Deep LearningComputer VisionNLPInformation RetrievalImage Processing
Highlights
  • Co-founded NextAura AI and builds its intelligent product capabilities.
  • Built FitCoach AI over one academic year.
  • Delivered a FastAPI house-price prediction endpoint during Ministry training.
  • Achieved 93.5% accuracy and 0.9354 macro F1 on the Cat vs Dog classifier.
  • Completed Anthropic AI training, created an Android manager, and graduated with a Very Good result.
  • Completed civic and leadership training through Al-Zamalah and Ana Osharek.
Certifications & training

Anthropic

  • Building with the Claude API
  • Claude Code 101
  • Introduction to Model Context Protocol
  • MCP: Advanced Topics
  • Introduction to Subagents
  • Claude on Amazon Bedrock
  • Claude on Google Cloud Vertex AI
  • Claude 101
  • AI Fluency Framework Foundations
  • AI Capabilities and Limitations
  • AI Fluency for Builders — Anthropic & CodePath

Cisco

  • Python Essentials 1
  • Python Essentials 2
  • Data Science Essentials with Python

AWS

  • AWS AI Practitioner Challenge

Udemy

  • Complete Computer Vision Bootcamp With PyTorch & TensorFlow
  • 100 Days of Code: The Complete Python Pro Bootcamp
  • Complete Android Development Bootcamp
Languages
Arabic
Native
English
Full professional proficiency
Volunteering

Al-Zamalah — Independent Election Commission & USAID Jordan

2024–2025

Completed 15 hours of structured training in dialogue, facilitation, critical thinking, and democratic participation.

Ana Osharek

2024–2025

Completed 15 hours focused on political participation, dialogue, and civic skills.

NEXAURA AI

Founder & AI Engineer · NextAura AI

LLM IntegrationRAG SystemsMCP & AI AgentsComputer VisionMachine LearningFastAPI & ML APIsAI ChatbotsPrompt EngineeringPythonFull-stack AIAndroid DevelopmentAWS AI ServicesAmazon BedrockGoogle Cloud Vertex AIClaude APIData AnalysisFeature EngineeringModel Evaluation

NEXAURA AI

Connect

Built in Jordan. Ready for the world.

  • AE
  • AF
  • AL
  • AR
  • AT
  • AU
  • BA
  • BD
  • BE
  • BG
  • BH
  • BR
  • CA
  • CD
  • CH
  • CI
  • CL
  • CM
  • CN
  • CO
  • CR
  • CV
  • CZ
  • DE
  • DJ
  • DK
  • DM
  • DZ
  • EC
  • EG
  • ES
  • FI
  • FR
  • GB
  • GE
  • GH
  • GN
  • GR
  • GT
  • HN
  • HR
  • HT
  • HU
  • ID
  • IE
  • IN
  • IQ
  • IR
  • IT
  • JO
  • JP
  • KP
  • KR
  • KW
  • LB
  • LK
  • LR
  • LY
  • MA
  • MC
  • ML
  • MR
  • MX
  • MY
  • NA
  • NE
  • NG
  • NL
  • NO
  • OM
  • PA
  • PE
  • PH
  • PK
  • PL
  • PS
  • PT
  • PY
  • QA
  • RO
  • RS
  • RU
  • SA
  • SD
  • SE
  • SG
  • SI
  • SK
  • SN
  • SV
  • SY
  • TD
  • TH
  • TN
  • TR
  • TW
  • UA
  • US
  • UY
  • UZ
  • VE
  • VN
  • YE
  • ZA