Architecting Production-Grade AI

Senior AI/ML Architect
& Full-Stack Lead.

With 12+ years of engineering foundation and 6+ years leading AI initiatives, I bridge the gap between research and production.I architect advanced AI ecosystems—spanning Deep Learning, Computer Vision, and Multi-Agentic RAG systems. By fusing state-of-the-art GenAI with robust backend engineering and Cloud (AWS/GCP) infrastructure, I turn experimental models into high-availability enterprise solutions.

#DistributedSystems#FullStackArch#RAG/Agents#ComputerVision#ProductionMLOps

System Architect

Designing fault-tolerant, scalable architectures that integrate Classical ML, Deep Learning, and Business Logic seamlessly.

Computer Vision & GenAI

Deep expertise in visual intelligence (OCR/Detection) combined with modern RAG and Agentic workflows.

Full-Stack Authority

I don't just hand off models. I build the API, the Database schema, the Container orchestration, and the Frontend UI.

AI Innovation built on
massive Full-Stack foundations.

I am a Principal AI/ML Engineer, but my roots lie in massive Backend Architecture and polished Frontend Engineering. With over 12 years of coding passion and 6+ years of professional execution, I bring a holistic software mindset to AI.

While my current focus is on Generative AI, RAG, and Computer Vision, this work is underpinned by years of building high-scale distributed systems. I don't just hand off model weights; I understand the database locks, the API latency, the Docker containerization, and the React state management required to serve them.

This unique combination allows me to own the entire delivery lifecycle. From optimizing PyTorch training loops to designing the user experience that consumes the prediction, I build intelligent systems that work in the real world.

6+
Years Professional

Leading AI & Engineering teams.

12+
Years Coding

Deep foundational expertise.

Technical Authority
System ArchitectureDistributed BackendFrontend/UXCloud NativeMLOps

Engineering Leadership

Massive Backend Scale

I architect for high-throughput and reliability. Whether it's optimizing vector indices for millions of documents or building event-driven microservices on Kafka, I ensure systems are resilient from Day 1.

Full-Stack Ownership

True ownership means understanding every layer. I move seamlessly between optimizing PyTorch models, tuning complex database queries, configuring K8s pods, and polishing frontend UI components in React.

Algorithmic Depth

I select the right tool for the job. From simple regression and decision trees (Classical ML) to complex Transformers and Vision systems, I prioritize efficiency and effectiveness over hype.

Technical Arsenal

A comprehensive toolkit for architecting modern, data-driven systems from the ground up.

AI / Machine Learning

PyTorchTensorFlowscikit-learnHugging FacePyTorch LightningComputer VisionNLP

GenAI & LLMs

RAGMulti-Agent SystemsLangChainPrompt EngineeringLoRA/QLoRAOpen Source LLMs

Backend & Full Stack

PythonFastAPIDjangoReactNext.jsTypeScript.NETC#

MLOps & Cloud

DockerKubernetesMLflowAirflowCI/CDGCP Vertex AIAzure MLAWS SageMaker

Mobile

FlutterDartFirebase

Work Experience

My professional journey in the tech industry.

AI/ML Engineer

Shory Insurance
Jun 2025 – Present Abu Dhabi, UAE (Onsite)
  • Designed scalable RAG pipelines (vector DBs, embeddings, prompt orchestration) for internal insurance workflows.
  • Built standardized RAG/agentic architecture: retrieval schemas, memory strategies, hybrid search routing, and evaluation frameworks.
  • Built ingestion/ETL pipelines for document chunking, metadata enrichment, embeddings, and continuous KB updates using Airflow.
  • Engineered OCR pipelines for RAG ingestion and document understanding (IDs/passports/fraud detection) using PaddleOCR and fine-tuned models.
  • Fine-tuned and deployed LLMs and Vision Transformers (LoRA/QLoRA/PEFT) with scalable serving via FastAPI and Docker.

AI/ML Engineer

eData Information
Jun 2023 – Jun 2025 Dubai, UAE (Onsite)
  • Developed a real-time vehicle valuation system using ensemble regression and deep learning techniques.
  • Created a VIN decoding system utilizing seq2seq models and ensemble learning.
  • Architected multi-agent/domain chatbots for insurance assistance & underwriting using RAG and fine-tuned LLMs.
  • Implemented MLOps pipelines using MLflow, Kubernetes, Docker, and Prometheus for observability.
  • Delivered solutions on GCP Vertex AI, Azure ML Studio, and Amazon SageMaker.

Software Engineer

Talal Tech
Feb 2022 – Jun 2023 Remote (Saudi Arabia)
  • Built high-performance APIs using Python, Django, and DRF.
  • Integrated full-stack solutions with React and Next.js frontends.
  • Established CI/CD pipelines with Docker and Kubernetes and implemented comprehensive testing frameworks.

Mobile Applications Developer

HoskaDev
Mar 2020 – Feb 2022 Remote (Algeria)
  • Developed Flutter mobile applications with complex state management.
  • Managed app store delivery and integration with 3rd-party services like Firebase, Google Maps, and Stripe.

Featured Projects

A selection of high-impact AI architectures, including open-source work and enterprise solutions.

Enterprise RAG & Document Intel

Case Study

A large-scale RAG system for processing insurance claims, automating 70% of manual document reviews.

Challenge

Handling noisy OCR data from scanned IDs and messy PDFs while maintaining high retrieval accuracy.

Outcome

Automated 70% of manual document review tasks, significantly speeding up claim processing.

PaddleOCRVector DBFastAPIAirflowLLMs

Multi-Agent Healthcare Assistant

An intelligent system utilizing RAG and autonomous agents to provide healthcare assistance and information retrieval.

Challenge

Orchestrating communication between multiple agents while maintaining context window efficiency.

Outcome

Reduced query response time by 40% compared to traditional search methods.

RAGLangChainDockerPythonReact

Vehicle Damage Vision System

Computer vision pipeline for detecting vehicle attributes and damages for automated valuation.

Challenge

Achieving high accuracy under varying lighting conditions and angles.

Outcome

Achieved 92% accuracy on validation set for vehicle make/model recognition.

Computer VisionOpenCVPyTorchPython

AI Vehicle Price Predictor

Real-time vehicle valuation engine using Deep Learning (TabNet) and Gradient Boosting. Analyzes market trends, vehicle condition, and mileage to predict precise resale values.

TensorFlowXGBoostFastAPIRedisBigQuery
Confidential / NDA Protected

VIN & Chassis Number Classifier

End-to-end OCR pipeline utilizing CRNN and Attention mechanisms to locate, rectify, and decode Chassis Numbers (VIN) from unconstrained, noisy vehicle images.

PyTorchCRNNYOLOv8OpenCVMobileNet
Confidential / NDA Protected

Patient Illness & Rx Predictor

Clinical decision support system predicting potential diagnoses and medication protocols based on patient symptoms and historical medical records.

BioBERTScikit-learnPostgreSQLDockerHIPAA-Compliant
Confidential / NDA Protected

Education & Certifications

Academic Background

MSc Artificial Intelligence

University of Science and Technology Oran, Algeria

2020 – 2022

Thesis: AI-based weather forecasting system for Algeria

BSc Computer Science

University of Science and Technology Oran, Algeria

2017 – 2020

Languages

Arabic (Native)
French (Fluent)
English (Fluent)

Professional Certifications

Certified Data Scientist & ML Professional

Dubai Knowledge / LIRC

DeepLearning.AI Machine Learning Specialization

Stanford / DeepLearning.AI

IBM AI Developer Specialization

IBM

NVIDIA Fundamentals of Deep Learning

NVIDIA

Get in Touch

Let's discuss how I can help build your next AI solution.

Contact Information

Chat with me