Expertise · MLOps and LLMOps

AI that keeps performing after launch

Putting a model or AI application live is only the start. Data shifts, users ask different questions, providers ship new model versions. MLOps makes sure you notice, and respond in a controlled way.

What MLOps and LLMOps cover

  • Evaluation as a test. Quality checks run in the CI pipeline, just like regular tests. No release without a passing evaluation.
  • Version everything. Models, prompts, data and configuration, so every outcome can be traced and rolled back.
  • Monitoring quality and cost. Not just whether the service is up, but whether the answers are still right and what they cost.
  • Controlled rollouts. New versions go to a share of users first, compared against the current version.
What that looks like in code
tests/test_eval.pyPython 3.12
import pytest

pytestmark = pytest.mark.asyncio
CASES = load_cases("eval/questions.jsonl")

# the evaluation set runs in CI, just like regular tests
@pytest.mark.parametrize("case", CASES, ids=lambda c: c.id)
async def test_answer_quality(case):
    answer = await assistant.ask(case.question)
    assert case.source_id in answer.sources
    assert judge.score(answer.text, case.expected) >= 0.8

# and we guard the cost per answer too
def test_cost_per_answer():
    assert costs.mean_per_answer(last_run()) < 0.02

Request profiles

Profiles of suitable MLOps engineers within 48 hours

  • No obligation: you only pay for hours worked
  • Dutch-speaking, from our own network and assessed by us
  • On-site, hybrid or remote

Indicative € 106–132 per hour, all-in, excl. VAT

Rather call? +31 35 7111 373

What we select for

  1. 01

    Automates

    Manual steps in a release are a reason to build something.

  2. 02

    Knows ML and infrastructure

    Understands what a model needs, and how to provide it on Kubernetes or in the cloud.

  3. 03

    Measures what matters

    Looks at the quality of outcomes, not just uptime.

Technology we often see

Tools change fast. We select on principles, but this is what the developers in our network use in practice.

MLflowKubernetesDockerKubeflowLangfuseWeights & BiasesTerraformGitHub Actions

Profiles we are looking for

Need a Python developer for this?

Send us your vacancy or request. Within 48 hours you hear which developers in our network fit and what it costs.