Production-oriented RAG architecture for enterprise question answering over PDF, DOCX, and JSONL sources — semantic chunking, multilingual embeddings, vector indexing, context validation, and grounded generation, served through a FastAPI REST API with fully local LLM inference.

Alaa Eddine Merzouk
Algiers, Algeria · Available for full-time roles
Retrieval, inference, and vision systems that ship — not just notebooks.
Computer Engineer specialized in Artificial Intelligence and Data Science, with hands-on experience developing Machine Learning, Deep Learning, Natural Language Processing (NLP), Retrieval-Augmented Generation (RAG), Large Language Models (LLMs), and Computer Vision solutions. Experienced in Python-based AI development, semantic search, vector databases, LLM inference, REST APIs, and containerized deployment using PyTorch, FastAPI, vLLM, Qdrant, Docker, Linux, and MLflow. Built a secure multilingual enterprise AI assistant using an on-premises architecture and developed computer vision systems for real-world detection and classification tasks. Currently expanding expertise in Cloud Computing, MLOps, and scalable AI deployment.
nearest-neighbour retrieval in 3D embedding space
Selected work
Projects across NLP, computer vision, mobile, and web. Pick a field to filter, or click an image to see it full size.
End-to-end wildfire detection pipeline built on a Kaggle dataset — training and validation with YOLO, evaluated on precision, recall, and mAP with a confusion matrix, plus detection examples and a FastAPI-served inference endpoint, containerized with Docker.
Real-time computer vision system detecting potholes and speed bumps in road imagery and video. Owned the full ML lifecycle — dataset prep, annotation, training, validation — with MLflow for experiment tracking.
Deep learning system that classifies brain MRI scans as healthy, glioma, or meningioma, with preprocessing, augmentation, a live upload demo, and a confidence score for each prediction.
Mobile platform for Algerian football players: book stadiums, find and join matches, and track player stats. Flutter client with clean, feature-first architecture on top of a Django REST backend with JWT authentication.
Multilingual Healthcare Booking Platform
Production booking site for the Algerian market in French, Arabic (full RTL), and English — CMS-driven content with Sanity, transactional booking flows with Supabase and Resend, and smooth Framer Motion interactions.
Velteix Agency Website
Agency website migrated from static HTML to Next.js, featuring a custom Three.js hero animation, 3D tilt cards, scroll-driven sections, animated page transitions, and a multilingual setup.
Experience
Artificial Intelligence Intern
Final Year Graduation Project (PFE)CASNOS — National Social Security Fund for Non-Salaried Workers · Algiers, Algeria
- Designed and built a secure multilingual RAG assistant for querying organizational documents in French, Arabic, and English.
- Built the full document pipeline: ingestion, preprocessing, semantic chunking, embeddings, vector search, context retrieval, and LLM response generation.
- Paired multilingual-e5-large embeddings with Qdrant for vector search and vLLM for local inference — no external LLM API in the loop.
- Exposed the system through a FastAPI backend designed for integration with internal enterprise applications.
- Tuned validation and retrieval to ~0.95 precision and a 0.05 hallucination rate, cutting manual document search effort.
- Ran the whole stack on-premises with Docker and Linux, keeping data in-house and removing recurring external API cost.
Freelance Web Developer
Velteix — Web Agency (Algerian & MENA market) · Remote, Algeria
- Built and maintained production Next.js sites for MENA-market clients, including a multilingual (FR/AR/EN) healthcare booking platform with RTL layout.
- Implemented CMS-driven content with Sanity, transactional booking flows with Supabase and Resend, and interactive front-end motion with Framer Motion and Three.js.
- Owned deployment and debugging on Vercel end to end, from initial build to production release.
IT Intern
ENNA — National Air Navigation Establishment · Algiers, Algeria
- Analyzed network infrastructure and information security practices inside a national air navigation organization.
- Investigated Linux-based systems, network services, and infrastructure monitoring in a live enterprise setting.
Education
Higher School of Computer Science (ESI) — Sidi Bel Abbès
State Engineer Degree in Computer Science, Artificial Intelligence & Data Science
Skills
AI & ML
Frameworks
LLM / RAG stack
Backend & Web
Data & Infra
Languages
Languages
Research interests
AI security, MLOps, cloud computing, distributed systems, large-scale data processing, computer networks.