KeykibotosTech consulting · Keykibotos
Data engineering, cloud and AI infrastructure for companies, by a staff engineer who has built them for some of France’s largest. What I build, five case studies with their figures, what people I worked with say, and how an engagement runs.
For companies whose data platform has outgrown the team running it. Remote, worldwide.
Book a 30-minute callOr write to me
Pipelines and data platforms, batch and streaming: ingestion, processing, lineage, and the monitoring that tells you when something breaks.
Infrastructure on AWS and Kubernetes, written as code, sized to what it carries, with its costs in view.
CI/CD, GitOps and deployments that are tested before they ship and can be rolled back.
Self-hosted open-source models in production: GPU inference, the pipelines around it, and what it costs to run.
DataOps engineer · 2025–2026
Pipeline deployment from weeks to hours
Deploying a data pipeline took weeks, and business teams waited on engineers for each one.
My part
I designed and implemented the production pipelines, and built the self-service data platform on Kubernetes, with GitOps deployments and the monitoring and alerting behind it.
What the team delivered
Staff software engineer · 2022–2023
API response time cut by 60%, at 1M+ requests a day
Rail operations data ran through a legacy monolith, with every system tied to the others.
My part
I architected its replacement: event-driven microservices in Scala on AWS, designed by domain.
What the team delivered
Data engineer · 2021–2022
2TB a day from 500K+ connected vehicles, turned into a dashboard for 200+ dealers
Telemetry from connected vehicles across Europe had to become something dealers could act on.
My part
I designed the Spark Streaming pipelines on AWS, implemented the predictive maintenance models, and created the fleet performance dashboard.
What the team delivered
Data engineer · 2019–2021
Batch processing from 8 hours to 45 minutes
Batch processing over 50M+ real estate listings took 8 hours.
My part
I architected the serverless data lake on AWS, developed the Spark jobs over it, built the REST APIs on top, and implemented the machine-learning pipeline for property valuation.
What the team delivered
Data engineer · 2017–2019
300M+ tracking events a day, ingested 10x faster
Parcel tracking produced a flood of events, to be ingested fast and in line with GDPR.
My part
I developed a generic ingestion platform on Kafka and Spark, and built it private by design.
What the team delivered
Roles and figures as on my LinkedIn profile.
I have had the pleasure of working with Ghazi at Le Groupe La Poste over the last few years. He is simply one of the most talented engineers I've ever worked with. He can solve any problem and he has a passion and drive to go way beyond what is expected in any situation. I would highly recommend Ghazi for Data engineering roles.
Je recommande vivement Ghazy qui a été un de mes consultants, le retour client était excellent. Ghazy est un consultant sérieux et engagé qui respecte les délais et les budgets et qui vous apportera son expertise pour la bonne déliverabilité de vos projets.
Translated from French: I strongly recommend Ghazy, who was one of my consultants; the client’s feedback was excellent. Ghazy is a serious, committed consultant who meets deadlines and budgets, and will bring you his expertise to deliver your projects well.
Very good freelance. I was his manager during his last mission and Ghazi Benahmed was perfect for his Scala/Spark + AWS mission. He build the whole Data AWS architecture, Terraform included 👌
Ghazi is a true senior data engineer with strong knowledges to resolve any kind of big data problem. It is a pleasure to work on a project with him and won't hesitate to work again with him if the future permits it. I deeply recommend him.
Two to three weeks
A written review of a platform: what it costs, where it is fragile, and what to change first. It ends with the report.
A fixed scope, over weeks or months
A platform, a pipeline or a migration, designed and delivered, with the documentation your team needs to run it.
A set number of days a month
Ongoing work on a platform: improvements, reviews and help when it matters.
No fixed packages. Each engagement is scoped after the call, agreed in a contract, invoiced, and paid by bank transfer.
I’m Ghazi Ben Ahmed, a staff engineer. Before founding Keykibotos, I spent more than fifteen years as a software and data engineer, mostly for the European market: most recently as a staff software engineer at SNCF and a DataOps engineer at EDF, and before that building data platforms for Stellantis, SeLoger, La Poste and Orange, and software for Société Générale.
Since 2025 I have also built AI infrastructure: a production platform on self-hosted open-source models, with its GPU inference.
HDP Certified Developer, and the Functional Programming in Scala specialization (EPFL).
Ghazi Ben Ahmed, founder · On LinkedIn
Tell me about your project here, or by email. I answer by email.