ARTIFICIAL
INTELLIGENCE
ENGINEER
My name is Adham Ehab, I'm Specialized in building production-grade Agentic RAG systems, distributed training pipelines, and scalable backend architectures. Experienced in orchestrating multi-agent workflows and optimizing LLMs for real-world business impact.
0+
PROJECTS DELIVERED
0+
UAE ENTERPRISE CLIENTS
UAE
RESIDENT Available for Hire
Full Stack AI Engineer
From 0 to 1: Architecting complete production apps with React, FastAPI, and Agentic AI integration.
40% ↓ Latency
50% ↓ Tokens
Multi-Agent Systems
MCP servers, RAG, and orchestration
TECHNICAL
ARSENAL
FEATURED
PROJECTS
Compass
AI Productivity System
Orchestrated domain-specific agents with a custom MCP server with Semantic Caching and dynamic tool retrieval, reducing latency to <100ms and cutting API costs by 30%.
Trace
RAG Powered FinTech Platform
Engineered a Hybrid RAG pipeline achieving 91%+ Context Precision (RAGAS). Built a low-cost ingestion pipeline using PaddleOCR and a Quantized SLM (Phi-3.5), reducing cost by 90%.
LiteDB
Lightweight Database GUI & AI Agent
Engineered a lightweight (5.4MB RAM) cross-platform database GUI using Tauri and Rust. Features a schema-aware Text-to-SQL Agent and on-device Vector Search using ONNX.
CAPTCHA-Solver
Two-Stage Computer Vision System
Engineered a two-stage vision system using YOLO11 to solve image-based CAPTCHAs with near 100% accuracy. Built a custom data engine generating 10k+ labeled synthetic samples.
SELECTED
WORK
WORK
EXPERIENCE
AI Engineer & Consultant
Engineered production-grade fulls-stack applications, agentic RAG systems and fine-tuned LLMs (Llama, Mistral) using LoRA. Developed an adaptive ML tracking system that reduced student dropout risks by 20% for client MC.
AI Engineer Intern
Secured 1st Place in a Kaggle competition (0.9440 score) by designing an end-to-end ML pipeline. Built distributed BERT training workflows on V-WLAN using Ray and optimized hyperparameters with Optuna.
AI & Robotics Intern
Deployed an on-premise Computer Vision attendance system with 95% face detection accuracy. Reduced manual administrative processing time by 90% via automated NLP query handling.