Building Intelligent
Systems
Autonomous agents, voice AI, computer vision, and MLOps infrastructure —
turning cutting-edge models into high-performance products.
AI & MLOps Stack
The tools I actually ship with — curated from real production projects.
Agent Frameworks
LLM & AI APIs
ML / Computer Vision
MLOps & Infra
Languages & Cloud
Frontend & Interfaces
Project Log
A timestamped record of autonomous agents, end-to-end AI products, computer vision architectures, and robust data systems.
Agents & LLMs
Agents & LLMs
Autonomous AI systems — sequential pipelines, voice agents, and RAG architectures that go far beyond simple API wrappers.
AI Stock Screening Agent
Sequential 4-phase autonomous agent: macroeconomic context → fundamental screening → technical triggers → catalyst validation. Delivers ranked shortlist with rationale.
Voice AI Invoice Generator
Voice AI assistant for generating invoices through natural language. Real-time turn detection, STT, TTS pipeline with LiveKit.
DesertVoice: Arabic RAG Agent
Voice-enabled customer support in Saudi dialect Arabic. Combines RAG with real-time audio using Gemini embeddings, Annoy ANNS, and LiveKit for low-latency streaming. Features dynamic knowledge base updates and Next.js frontend with waveform visualization.
Arabic Customer Support RAG Voice Agent
Sophisticated RAG voice agent for Arabic customer support (Saudi dialect). Voice-first interaction via LiveKit, vector search with Annoy for context retrieval, Gemini text-embedding-004 for 768-dim embeddings. Features a search_support_docs function-calling tool and dynamic knowledge base updates via file upload.
End-to-End AI Products
End-to-End AI Products
Production-ready systems where AI is deeply wired into the product's core logic. From geometric computer vision pipelines to multi-modal content generation.
FloorPlan Pro 3D
Full-stack 2D→3D floor plan converter. AI detects walls, doors, windows via Roboflow. Three.js interactive 3D editor.
AI Exam Management System
Mistral AI integration for intelligent question generation from OCR'd PDF/Word docs. Multi-mode: AI, Question Bank, Manual. Arabic RTL support.
AI Medical Competency Manager
Manages medical staff competencies with Mistral AI analysis. Full-stack with Drizzle ORM and comprehensive database tracking.
Neuro-Canvas: Visual NN Architect
AI-powered visual code generator for neural networks. Drag-and-drop model building with PyTorch/TensorFlow code export.
Computer Vision
Computer Vision
Building systems that see and understand — from semantic image search to real-time object detection in production pipelines.
VisualVoyager: Advanced Visual Search Engine
High-performance image search integrating state-of-the-art CV models. Two-stage pipeline: CLIP + SuperPoint global features → FAISS ANN index, then ORB + RANSAC geometric reranking. Also supports Meta's Perception Encoder (PE-Core-L14-336) with smart directory caching and diversity re-ranking.
Systems & APIs
Systems & APIs
From robust data pipelines to real-time dashboards and scalable distribution layers — the engineering backbone that keeps AI products running flawlessly.
ResponsePlus Medical Dashboard
Interactive dashboard for visualizing employee data, locations, and project assignments across regions for a medical services company.
Pharmacy Management System
Core pharmacy platform: RESTful API, inventory management database, and mobile integration support.
Radaar El Dawa: Mobile App
Mobile app for searching drug availability, placing orders, and real-time pharmacy communication.
About>Who am I?
I engineer AI systems that work in the real world. My background spans autonomous agents, voice pipelines, computer vision, and scalable ML infrastructures. I care obsessively about the gap between a fragile notebook demo and a resilient deployed product — and I close it.
Ready to Build?
Looking to bring AI into your product? Let's talk about agents, pipelines, or anything in between.
Pick a time that works for you
© 2026 · Built with Next.js, Three.js, LangGraph vibes & Framer Motion