Avatar

Aymen Elghali

aymenelghali3@gmail.com

24 yrs old
Target Roles
Frontend DeveloperBackend DeveloperMobile App DeveloperData ScientistAI / ML Engineer

Share Profile
Project Demo & Video Introduction

Professional Summary

I’m a software engineer from Tunisia with an engineering degree in Data Engineering from ENET’Com Sfax. For the last year and a half, I’ve built production software at two European AI companies, qINSIGHT in Belgium and Sagy AI in the Netherlands.

I build backend APIs and services in Python and TypeScript, design AI agent and RAG workflows, and deploy with Docker and Azure. I enjoy owning work end to end and debugging across systems. I speak English, French and Arabic, and I’m available full-time now.

Technical Skills

HTML/CSSJavaScriptPHPPythonNext.jsMySQL / PostgreSQLREST APIs / GraphQLGit / GitHubDocker / KubernetesReact Native / FlutterC++ / C#

Education

Master's Degree

Experience

0-1 Years (Fresher)

Notable Projects

LLM document analysis platform (qINSIGHT, in production). I built AI agent workflows with the OpenAI Agents SDK and RAG pipelines that analyse documents and transcripts. Every citation the model gives is checked against the source before it reaches the user. I also redesigned the large-file transcription pipeline to remove repeated file transfers between the browser, backend and Azure Blob Storage, and added async processing with Redis progress tracking. Stack: Python, TypeScript, React, PostgreSQL, Redis, Docker, Azure.
Enterprise knowledge platform (Sagy AI). I shipped full-stack features with Django REST APIs and React, plus NLP pipelines for classification and information extraction. I fine-tuned DistilBERT and built a lighter TF-IDF + Logistic Regression alternative to reduce inference cost. Stack: Python, Django, React, Docker, MLflow.
Natural-language data assistant (Proxym IT). I built a multi-agent system with LangChain and Mistral-7B that turns plain-language questions into SQL and Elasticsearch queries, with validation, caching and logging so the results are reliable and auditable.