Bader Zakaria
AI & Agentic Systems Engineer
Agents · RAG · Knowledge Graphs · Data Platforms · AI Products
I design intelligent systems that retrieve, reason, use tools and operate inside real business workflows — while staying observable, secure and controllable.
Tap a domain
Every figure links to its source.
28
countries
Mobility data platform built from the ground up to 28 countries
Source: Smart Mobility Lab~10×
data volume
Approximate data-volume increase handled by the platform
Source: Smart Mobility Lab7+
years experience
Professional engineering experience — from software to agentic AI systems, since 2019
Source: Intelligile3
working languages
French · English · Arabic — and products built in all three
Source: Multilingual Knowledge Platform02Work
Case studies written for three readers: skim the thesis, read the system, inspect the decisions.
Data platform · Smart Mobility Lab
Production data and AI pipelines turning fragmented national vehicle-registration sources into one trusted reference — built from the ground up to 28 countries.
“Experimentation is useful. Production needs repeatable contracts.”
Specialised agents search the web for vehicle prices, validate them, normalise currencies and persist the result — with provenance — in a Neo4j knowledge graph.
Human-in-the-loop · Deterministic control plane
Agents where useful. Deterministic software where mistakes have consequences.
A Telegram-first control plane for researching, drafting, approving and publishing financial content. AI drafts; a state machine decides; a human approves.
A private French-, Arabic- and English-language document library with personal libraries, search, recommendations and an administrative publishing workflow.
ML foundation · Utrender · 2021
A Python/OpenCV system measuring how people interact with advertising screens, with an associated IoT web application — AI work that predates the LLM wave.
“Generative AI is one more capability — not a replacement for the fundamentals.”
03How I think about production AI
01
What are we solving, for whom — and what happens when the system is wrong?
02
What is probabilistic, what stays deterministic, and which tools may an agent touch?
03
Models, retrieval, tools, APIs, data and interfaces become one architecture.
04
Quality needs evidence: tests, validation rules, failure analysis, feedback.
05
Sensitive actions need permissions, audit trails and, where it matters, a human.
06
Logs, monitoring, cost and deployment decide whether it survives outside a demo.
04Accumulated capability