I'm a freelance AI platform engineer, specialized in LLMOps, AIOps and MLOps. I design and operate reliable, scalable and observable AI platforms on Kubernetes and the cloud - from the data pipeline to serving LLMs in production.
IASWITCH is my freelance practice, specialized in AI platforms. I help engineering teams and companies industrialize AI: laying down cloud-native foundations, automating ML/LLM pipelines, and making sure systems stay reliable once they hit production.
My approach is pragmatic and platform-first: I build golden paths your teams can reuse, with automation, observability and security baked in from day one - not bolted on afterwards.
From the underlying platform to operating the models, I cover the full lifecycle.
Deploying, versioning and monitoring LLMs in production: RAG, fine-tuning, evaluation, cost and latency control.
Anomaly detection, alert correlation and automated remediation to cut noise and MTTR across your operations.
Internal developer platforms (IDP) and golden paths on Kubernetes so your developers ship faster and safely.
End-to-end pipelines: feature stores, reproducible training, model registry, deployment and drift monitoring.
Reproducible cloud-native architectures on AWS/GCP, provisioned with Terraform, GitOps CI/CD and cost control.
SRE, SLO/SLI, secrets management and supply-chain security for robust, compliant AI platforms.
Designed an auto-scaling inference platform serving multiple models with smart routing, cutting GPU costs by 40%.
Secure RAG pipeline over internal data with continuous evaluation, guardrails and full answer traceability.
Self-service golden paths on Kubernetes: from commit to prod in minutes, with observability and security by default.
Alert correlation and anomaly detection to cut MTTR by 3x and remove 70% of non-actionable alerts.
I frame your goals, constraints and the current state of your infra and data.
I propose a pragmatic target architecture, scoped and documented, with a phased plan.
Incremental delivery in IaC and GitOps - tested, observable and handed over to your teams.
Go-live, monitoring, cost optimization and upskilling your team.
LLMOps is the set of practices for deploying, versioning, monitoring and evolving large language models (LLMs) in production: RAG, fine-tuning, evaluation, cost and latency control.
MLOps industrializes ML models (training, registry, deployment, drift). LLMOps adds LLM-specific concerns (prompts, RAG, evaluation). AIOps applies ML to IT operations to cut alert noise and MTTR.
I deploy LLMs via vLLM on Kubernetes with per-request GPU autoscaling, multi-model routing, guardrails and observability (cost, latency, quality), provisioned as Infrastructure as Code.
IASWITCH is a freelance AI platform engineer based in Belgium, working remotely across Europe on LLMOps, AIOps, MLOps and Platform Engineering engagements as an independent consultant.
Let's talk. Tell me about your context and goals, and I'll get back to you quickly with a first read.