Jerome Renaux

Fractional senior platform engineer

What follows is a short selection of pain points I have solved before that I think are relevant for 3Ai.

Selected relevant experience

01

AI service launch path

Pain point
New model-backed services need consistent packaging, deployment, observability, and ownership before customer traffic arrives.
How I tend to approach it
Define a launch template with service scaffolding, deploy gates, SLO defaults, dashboards, and runbook prompts.
Useful result
Teams can move from prototype to operated service without rebuilding platform decisions each time.
02

Evaluation and release workflows

Pain point
Model and prompt changes can ship through unclear approval paths, making quality regressions harder to catch.
How I tend to approach it
Create a release workflow that ties eval results, artifact versions, rollout controls, and incident signals together.
Useful result
Product teams get faster iteration with a clearer line between experiment, staged rollout, and production change.
03

Cost-aware inference platform

Pain point
Inference cost and capacity can drift as usage patterns, providers, and model choices evolve.
How I tend to approach it
Expose unit-cost reporting, workload tagging, autoscaling defaults, and provider-aware runtime options.
Useful result
Engineering can discuss performance, quality, and margin using shared operational data.

Where I fit

Good fit if:

  • You have 4 to 25 engineers.
  • Your backend or product engineers are spending noticeable time on infra or deployment issues.
  • You are not ready to hire a full-time platform engineer.
  • You need someone senior enough to diagnose, prioritize, and implement.
  • You want pragmatic execution, not an abstract DevOps transformation.

Bad fit if:

  • You need full-time availability.
  • You mainly want ticket-taking DevOps labor.
  • You need 24/7 incident response ownership.
  • You are looking for the cheapest possible implementation capacity.

Typical ways to work together

Schedule a call