Mohamed Aziz BEN HAHA
Open to missions & training · France and Tunisia
I build AI systems & cloud infrastructure
Cloud & DevOps engineer (CKA, AWS Solutions Architect & SysOps), AI/ML practitioner, trainer with 1,000+ hours delivered to universities and companies, and PhD student in machine learning for embedded IoT at Sup'Com × École Centrale de Lyon.
Taught at, built for, studied with
AI systems and the infrastructure they run on. RAG pipelines that answer from your own documents with verified citations, machine-learning models trained on real data and deployed on the cloud or on-premise, and the automation around them: CI/CD, infrastructure as code, monitoring, high availability. In production at Ooredoo, PASS and IPACT.
Cloud, DevOps and AI, end to end. Learners go from zero to a working AI/ML model deployed on cloud infrastructure or locally, using managed cloud AI services or models they build themselves.
PhD in machine learning for embedded IoT: an autonomous platform that watches water quality in real time.
End-of-studies (PFE) and end-of-year projects for engineering and master's students, at university or inside companies: from the subject and architecture to the defence.
Natural language in, validated actions out
My personal operating system, on my phone every day. It started as a gym log and keeps growing: training, notes, tasks, reminders, calendar, groceries, and next, money. Type or speak one sentence and an AI turns it into validated actions, with one-tap undo. Each module is its own app; they share one sync key and read each other's data, never write it.
A personal AI that improves itself
A self-evolving personal AI that runs on my own machine. A small local model does the thinking; it remembers across sessions, recalls what it has learned with its source, wanders the web on its own and leaves me notes, and can reshape a bounded part of its own code — kept only if the tests still pass, or a guardian restores the last version. A few core principles it can never rewrite, and a kill switch only I hold.
Tools for the whole PhD journey, from subject to defence
Four connected tools for students and supervisors: find a subject and a supervisor, publish the thesis, question your own sources, and run the research itself. Microservices deployed on Kubernetes with Helm.
Autonomous, intelligent water-quality biomonitoring
Microbial fuel cells used as living sensors, an IoT acquisition box, and machine learning that runs on the device to read water quality in real time. Franco-Tunisian project (PHC Utique 2025–2027) between Innov'COM, Sup'Com, and the Ampère lab, École Centrale de Lyon.
YouTube channel and podcast on research
“The Researcher's Path: From Idea to Thesis”: episodes and shorts that walk researchers from the first idea to the defence, from choosing a topic and framing the problem to methodology, writing and analysis.
Life OS
My training log, used at every session: programme, logging, progress charts, an AI coach and cloud sync. It also feeds its food list to Life OS groceries.
Notes inside Life OS, written and edited from the same one-sentence AI bar as tasks and reminders.
In progress · spending and budget tracking, the next Life OS module, on the same sync and AI action bar.
Thesis ecosystem
The PhD library, with a page for every researcher: manuscript, articles, presentation, images, videos and the team behind each thesis. Open source.
Supervisors propose thesis subjects; students apply or propose their own. Applications are tracked from pending to accepted.
A free Windows app for students: point it at a folder, ask questions, get answers that cite your own files, page and passage, checked sentence by sentence. Local models, cloud optional.
The research system I run my PhD on: one workspace per paper with literature sync, concept notes, citation audit and a peer-review pass, a lab notebook with write-once data, and the annual report.
GreenWaterGuard
ESP32 firmware reading pH, ORP, conductivity, dissolved oxygen and temperature probes over Modbus RTU and two ADCs, publishing JSON over MQTT. Field setup through a captive Wi-Fi web app.
Parametric, code-generated CAD for a floating microbial-fuel-cell sensor unit: three prototypes, buoyancy checks, print plates for a Bambu printer, and an automated test suite.
Programmes I design and deliver for universities, companies and public employment programmes. Open a course to see its outline and preview the full programme, or ask for a version adapted to your team: on-site in France or Tunisia, or remote.
Published2026
M. A. BEN HAHA, A. Bohli, N. Haddour, R. Bouallegue
Electronics 15(11):2240 · MDPI · DOI 10.3390/electronics15112240
Under review2026
M. A. BEN HAHA, A. Bohli, N. Haddour, R. Bouallegue
Journal of Hydroinformatics · IWA Publishing
In preparation
Working title · GreenWaterGuard PhD
In preparation
Working title · GreenWaterGuard PhD
Build it. Automate it. Then teach what runs in production.