I am a staff engineer with more than fifteen years of experience in software engineering, data engineering, AI infrastructure, and high-performance data platforms.
I build production systems across Python, Scala, PyTorch, Spark, Kafka, Kubernetes, AWS, and Terraform. I apply test-driven development throughout the engineering lifecycle and focus on reliable, measurable operational outcomes.
I have delivered large-scale platforms for organizations including SNCF, EDF, Stellantis, SeLoger, La Poste, Orange, SUEZ, and Société Générale CIB. My work has included event-driven services, streaming pipelines, serverless data lakes, DataOps platforms, and machine-learning pipelines.
I specialize in performance and cost optimization. I have reduced batch processing from eight hours to forty-five minutes, optimized streaming platforms processing billions of events, reduced cloud and processing costs, and improved API response times at high request volumes.
In 2026, I built a production AI content-generation platform using self-hosted open-source models. I trained LoRA adapters for Wan2 video generation, deployed GPU inference on Runpod, and optimized GPU usage, inference configuration, and training workloads to lower costs.
I am also the founder of Keykibotos, where I provide data, cloud, and AI infrastructure consulting while building digital products based on open data.
I founded Keykibotos in Singapore to provide consulting in data, cloud, and AI infrastructure.
I also develop digital products built on open data, combining production engineering, cloud architecture, and AI platform expertise.
I designed and developed a production-ready, test-driven AI content-generation platform using self-hosted open-source models. I built the end-to-end pipeline for automated book and animated-video generation.
I trained LoRA adapters on Wan2 using PyTorch and Diffusers, and deployed GPU inference workloads on Runpod. I optimized CUDA environments, moved FFmpeg encoding and decoding from GPU to CPU, tuned resolutions and frame counts, right-sized GPU instances, reduced training steps, and added cost-per-run metrics and spend alerts.
I built a self-service data platform on Kubernetes using GitOps deployments. Pipeline deployment time was reduced from weeks to hours, enabling business teams to deploy their own pipelines.
I designed and implemented more than 15 production pipelines processing 200 GB per day with 99.9% uptime, including monitoring and alerting. The team reduced infrastructure costs by 40%.
I developed an ESP32 monitoring and control board for a hydroponic farm.
My work covered sensors, embedded monitoring and control capabilities, and cloud workflows.
I architected the replacement of a legacy monolith with event-driven microservices in Scala on AWS MSK. The platform processed more than 500 GB of rail operations data per day and was migrated with zero downtime.
The team reduced API response time by 60% at more than one million requests per day and reduced AWS costs by €200K annually.
I designed Spark Streaming pipelines handling 2 TB per day from more than 500,000 connected vehicles. The platform processed over 10 million events per hour with sub-second latency.
I reduced processing costs by 45% through partitioning and cluster optimization. I also implemented predictive-maintenance models that identified issues two weeks before failure with 87% accuracy.
I developed more than 30 optimized Spark jobs for over 50 million listings, reducing batch processing time from eight hours to forty-five minutes.
I architected a serverless AWS data lake, built REST APIs with p99 latency below 100 ms at 500,000 requests per day, and implemented a property-valuation ML pipeline that improved accuracy by 35%.
I developed a generic Kafka and Spark ingestion platform for more than 300 million tracking events per day. The platform increased ingestion speed tenfold and reduced storage costs by 60% through compression and partitioning.
I designed the platform with privacy and GDPR requirements in mind while maintaining query performance. I also implemented recursive schema processing using Matryoshka, Scala, and functional programming.
I worked as a Data Engineer in Paris.
Detailed responsibilities and project outcomes were not provided in the resume.
I worked as a Software Developer in Paris.
Detailed responsibilities and project outcomes were not provided in the resume.
I worked as a Software Developer in Paris.
Detailed responsibilities and project outcomes were not provided in the resume.
I worked on application architecture, API management, code review, test-driven development, behavior-driven development, continuous integration, and team management.
I contributed to engineering practices and delivery processes across software projects.
I worked as a Full Stack Developer in Tunis.
Detailed responsibilities and project outcomes were not provided in the resume.
Engineer’s degree, Computer Science
Engineer’s degree in Computer Science.
Licentiate degree, Theoretical and Mathematical Physics
Licentiate degree in Theoretical and Mathematical Physics.
This professional hasn’t added portfolio projects yet.
This professional hasn’t listed any services yet.
