ak:Abderrahman
Kouki
FR/EN
CV

+ AI & software engineering

From complex data
to useful products.

I’m Abderrahman Kouki. I build AI systems, reproducible data workflows, and software that solves everyday business problems.

Explore my work Recruiter brief

Based in Tunis · French & English C1

ABK / PORTFOLIO2026
Abderrahman Kouki
Applied AI. Practical engineering.
85%less payroll processing workloadTimeclock automation · DS Informatique
End to endfrom requirements to supportClient delivery · Kouki Software
Reproducibledata, training selections & deploymentsMLOps platform · Anavid

01 / Selected work

The work behind
the skill set.

Professional experience first. A closer look at what I contributed, the systems I built, and the tools I used.

01MLOps / Internship
D→ML
01Visual / action data02Versioned datasets03Training selections

Reproducibility across the data lifecycle

AnavidFeb — Jun 2026

A traceable path from raw data to training datasets.

An MLOps platform for visual and action-detection data, connecting KPI dashboards, dataset versioning, and reproducible training selections.

PythonStreamlitDVCOVH S3DockerGitHub ActionsParquet
Read project details

The context

Work with visual and action-detection datasets while keeping track of the data selected for training.

My contribution

  • Built KPI dashboards and data traceability workflows.
  • Automated S3 manifests, dataset versioning, and reproducible training selections.
  • Automated deployment workflows with GitHub Actions and Docker/GHCR.

What this demonstrates

A practical foundation in data lineage, reproducibility, and the software around machine learning.

02Automation & RAG / Internship
−85%
01Timeclock records02Automated integration03Payroll workflow

85% less payroll processing workload

DS InformatiqueJul — Aug 2025

Less manual processing. More usable business information.

Factory timeclock integration reduced payroll processing workload by 85%. A separate RAG system summarized internal documents and business reports.

PythonSQLn8nQdrantRAG
Read project details

The context

Reduce manual work around factory timekeeping and help teams use information in internal reports.

My contribution

  • Automated the integration of factory timeclock data.
  • Connected business workflows using Python, SQL, and n8n.
  • Built a RAG system with Qdrant for report and document summaries.

What this demonstrates

Experience turning a specific operational problem into automation and an AI-assisted information workflow.

03Business software / Founder
K/S
01Client requirements02Custom software03Deployment & support

Ownership throughout the client lifecycle

Kouki Software2020 — 2026

Software built around the way a business actually works.

Custom accounting, tax compliance, and ERP solutions for local businesses, with responsibility for the full client lifecycle.

PythonWinDevSQLXMLLangChain
Read project details

The context

Understand the requirements of local businesses and turn them into tailored management software.

My contribution

  • Founded and led a software company delivering accounting, tax compliance, and ERP solutions.
  • Gathered requirements directly with clients.
  • Owned implementation, deployment, and ongoing support.

What this demonstrates

Client-facing delivery experience: translating a business need into software and staying responsible after launch.

Additional projects

A space to build
and explore.

Explore my GitHub
04

Multi-agent debate

AI-versus-AI and human-versus-AI debate modes, expert personas, and real-time text-to-speech.

LangGraph · CrewAI · Ollama
05

Gym management SaaS

Booking automation, row-level security, and real-time updates in a full-stack management application.

React · Supabase
06

MLOps automation

Pipeline orchestration, data versioning, experiment tracking, and deployment through CI/CD.

ZenML · MLflow · DVC · Kubernetes
07

Diabetes risk tool

Distributed preparation of CDC BRFSS data and a deployed prediction interface. An academic ML project.

Spark · Hadoop · XGBoost · Streamlit

02 / Experience

Technical depth.
Business context.

Interested in AI Solutions, Applied AI, MLOps and full-stack roles.

Feb — Jun 2026

MLOps intern

Anavid · Paris / Remote

Data traceability, reproducible datasets, dashboards, and automated deployments.

Jul — Aug 2025

Python developer intern

DS Informatique · Remote

Factory timeclock integration and retrieval-augmented document summaries.

Jul 2023

Accounting & IT intern

Bureau Moncef Zayani · Tunisia

Compliance reports and database connectivity for financial tools.

2020 — 2026

Founder & lead developer

Kouki Software · Tunisia

Custom business software from discovery to deployment and support.

03 / Background

Mathematical foundations.
A builder’s mindset.

My background combines applied mathematics with software engineering, Data Science, and AI. I enjoy the full path from understanding a problem to building something people can use.

Languages

Arabic — native · English — C1 · French — C1 · German — basic

Selected certifications

Microsoft Azure AI-900 · NVIDIA Deep Learning, NLP & Diffusion · PMI Management Ready · Cisco CCNA

Education

2023 — 2026

École Polytechnique de Sousse

Software engineering · Data Science & AI specialization

2021 — 2023

Université Paris Dauphine — PSL

Bachelor in applied mathematics for computing & decision sciences

2018 — 2021

IPEIT Tunis

Mathematics & physics preparatory classes

04 / Recruiter brief

A few useful answers.
Before we meet.

A quick overview grounded in my experience and projects.

Which roles are you interested in?

AI Solutions Engineer, Applied AI Engineer, MLOps, and full-stack roles where I can connect technical implementation with business needs.

See relevant work
What is your experience with clients?

At Kouki Software, I handled requirements gathering, implementation, deployment, and support for local businesses. That experience informs how I approach solution engineering.

Client delivery
Where have you worked with RAG and agents?

I built a RAG system for business reports at DS Informatique, and explored multi-agent interactions in a debate game using LangGraph, CrewAI, and Ollama.

RAG experience
What is your MLOps experience?

My Anavid internship focused on dataset traceability, reproducible training selections, and automated workflows. My MLOps automation project used ZenML, MLflow, DVC, and Kubernetes.

MLOps experience

Let’s connect

Have a problem
worth solving?

I’d love to hear about your team, the product you’re building, and where I could contribute.

Email me Download CV