ClearML Setup for MLOps and Experiment Tracking

Lost experiment context? A proper ClearML MLOps setup solves this at the protocol level: every run captures everything — from git commit to GPU consumption. In a typical ML project, a data scientist spends up to 30% of their time retraining; ClearML cuts that time by 70%. Our team has 10+ years of MLOps experience, and we guarantee a reproducible setup.

In practice: a team of five specialists can spend hours searching for the right configuration. ClearML provides automatic experiment tracking, data management, and orchestration. We have deployed the platform in 50+ projects, from startups to enterprise. In one case — a fraud detection startup — reproduction time dropped from 3 days to 4 hours. Contact us for a free consultation and project assessment.

Why ClearML is Better than MLflow for Teams

Both tools solve tracking, but ClearML wins with its built-in task queue and automatic reproduction. The ClearML Agent can execute any experiment on another machine with a single command. MLflow requires external schedulers (Airflow, Kubeflow) for reruns. In our projects, ClearML speeds up reproduction 10x, saving significant GPU idle costs.

How to Set Up ClearML on Your Own Server

Installing the Self-Hosted Server

# Docker Compose for self-hosted
git clone https://github.com/allegroai/clearml-server
cd clearml-server
docker compose -f docker-compose.yml up -d

# Web UI: http://localhost:8080

For production, add PostgreSQL and MinIO/S3 via environment variables. We have prepared a docker-compose.prod.yml template that includes SSL and backups. Setup on GPU nodes requires images with CUDA and NVIDIA drivers. The Agent automatically detects GPUs — not a single extra line of code.

Basic Usage

from clearml import Task, Logger

# Initialization — automatically captures git status, pip packages, config
task = Task.init(
    project_name="Fraud Detection",
    task_name="LGBM Baseline",
    task_type=Task.TaskTypes.training,
)

# Parameters
task.connect({
    "learning_rate": 0.05,
    "n_estimators": 500,
    "dataset_version": "v2.3"
})

# Metric logging
logger = task.get_logger()
for epoch in range(100):
    logger.report_scalar("Loss", "train", iteration=epoch, value=train_loss)
    logger.report_scalar("Loss", "val", iteration=epoch, value=val_loss)
    logger.report_scalar("F1", "val", iteration=epoch, value=val_f1)

# Tables and images
logger.report_table("Test Predictions", "Confusion Matrix", iteration=0, table_plot=cm_df)
logger.report_matplotlib_figure("ROC Curve", "ROC", iteration=0, figure=fig)

All metrics are available in the web UI immediately after launch. No additional setup required. You can log not only scalars but also images, tables, and audio.

ClearML Agent for Reproduction

Unique feature: automatic reproduction of any experiment:

# Start the agent (on another machine, including GPU)
clearml-agent daemon --queue default --detached

# Clone and re-run an experiment
clearml-agent execute --id <task_id>

The Agent pulls the environment from cache. Our clients save up to 30% of time on retraining.

Hyperparameter Optimization

from clearml.automation import HyperParameterOptimizer, RandomSearch

optimizer = HyperParameterOptimizer(
    base_task_id=task.id,
    hyper_parameters=[
        UniformParameterRange("learning_rate", min_value=0.001, max_value=0.1),
        DiscreteParameterRange("n_estimators", values=[100, 200, 500]),
    ],
    objective_metric_title="F1",
    objective_metric_series="val",
    objective_metric_sign="max",
    max_number_of_concurrent_tasks=4,
    optimizer_class=RandomSearch,
    total_max_jobs=50,
)
optimizer.start()

Hyperparameter optimization runs on the queue without blocking — agents execute tasks in parallel. Results are immediately available in the comparison table.

Common Mistakes When Setting Up ClearML

Additionally: many forget to configure artifact retention policies — by default ClearML stores everything indefinitely, filling up storage. We recommend setting a TTL via server configuration.

What's Included in Our Work

We provide a complete package:

• ClearML deployment on your infrastructure (Docker, Kubernetes, bare metal)
• Integration with your tools (GitLab, Jupyter, S3, GPU)
• Agent and queue configuration
• Usage documentation
• Team training (2–4 hour workshop)
• Support during launch phase (2 weeks)

Our team has over 10 years of MLOps experience and we guarantee a reproducible setup. We are a certified ClearML implementation partner.

MLOps — Wikipedia — terminology and methodologies.

Our Process

1. Analysis: audit of current MLOps stack, scalability requirements
2. Design: server, queue, and storage architecture
3. Implementation: deployment, integration setup, task template creation
4. Testing: reproducibility checks, load testing
5. Deployment: handover to operations, team training

Our Results

Over several years, we have implemented ClearML in 50+ projects, from startups to enterprise. Average reduction in experiment reproduction time: 70%. Clients save up to 50% of their MLOps infrastructure budget when migrating from MLflow Enterprise. We have a 95% success rate in reducing reproduction time.

Order ClearML setup for your stack. We will assess your project for free — contact us. Get a consultation today.

MLOps: Infrastructure for Training, Deploying, and Monitoring ML Models

The model is trained, metrics — F1 0.94 on validation. Three months later in production, quality drops by 12%. No one knows when — there is no monitoring. It's impossible to retrain quickly — the training script is in a Jupyter notebook of a data scientist who has already left. Data for retraining is collected manually from three disparate systems. About half of the projects come to us with this pain. We build a turnkey MLOps platform: from experiment tracking to automatic deployment and data drift monitoring. We will assess your infrastructure in 1–2 weeks, and in 4–6 weeks you will get a basic MLOps core running in production. Our team has 10+ years of experience in ML infrastructure, over 50 implementations.

How does MLOps infrastructure benefit your ML projects?

Experiment Tracking and Reproducibility

Without tracking, an ML project turns into chaos: it's unclear which checkpoint is better, which hyperparameters were used, which dataset. Reproducing a result a month later is a quest.

Why is experiment tracking the foundation of reproducibility?

MLflow is an open source standard for tracking. It logs parameters, metrics, artifacts (models, graphs), and code. MLflow Model Registry is a centralized model storage with versioning and lifecycle stages (Staging → Production → Archived). Deployment via MLflow Serving or integration with external systems.

Typical initialization in code:

import mlflow

mlflow.set_experiment("fraud-detection-v2")
with mlflow.start_run():
    mlflow.log_params({"learning_rate": 3e-4, "batch_size": 64, "epochs": 10})
    mlflow.log_metric("val_f1", val_f1, step=epoch)
    mlflow.pytorch.log_model(model, "model")

This is the minimum. In production, we add logging of system metrics (GPU utilization, memory), dataset (hash, version), code (git commit hash). Weights & Biases — richer UI, collaboration features, sweep for hyperparameter optimization. MLflow — for on-premise deployment without external dependencies.

DVC (Data Version Control) — versioning of data and models on top of git. Data is stored in S3/GCS/Azure Blob, only metadata (hashes) in git. dvc repro reproduces the entire pipeline from raw data to metrics.

To ensure reproducibility of training, fix random seeds (torch.manual_seed, numpy.random.seed, random.seed) and record them in experiment metadata. Without this, debugging irregular results is painful. Log the dataset version (DVC hash) and git commit — then any experiment can be reproduced down to the byte.

Pipeline Orchestration: Kubeflow, Airflow, Prefect

A pipeline orchestrator becomes necessary when: A 100-line training script in cron is fine for simple tasks. But as soon as you have a multi-step pipeline (data loading → preprocessing → feature engineering → training → validation → deployment if quality above threshold), you need an orchestrator with retry logic, visualization, and alerts.

Kubeflow — Kubernetes-native orchestrator for ML (see Kubeflow). Each step is a Docker container. Supports parallel steps, conditional branches, artifacts between steps. Integrates with Katib (AutoML), KServe (serving), Feast (feature store).

Apache Airflow — more general DAG orchestrator. Wide ecosystem of operators (S3, Spark, DBT, Kubernetes). Easier to deploy if Airflow already exists in the company.

Prefect / Metaflow — less boilerplate. Prefect 2.x with @flow and @task decorators — quick start for small teams.

Typical training pipeline architecture on Kubeflow:

1. Data ingestion component — fetches data from S3/DB, validates schema via Great Expectations
2. Preprocessing component — transformations, normalization, train/val/test split
3. Training component — training on GPU, logging to MLflow
4. Evaluation component — metric calculation, comparison with baseline in Model Registry
5. Conditional deployment — deploy only if new model is better than current by >2% F1

Each component is a separate Docker image. Pipeline is versioned in git. Scheduled run (retraining once a week on new data) or manual.

Model Registry and Lifecycle Management

Model Registry is not just a checkpoint store. It is a centralized system that knows:

• Which model is currently in production (and with what metrics)
• History of all versions with training parameters
• Metadata: dataset, git commit, validation results
• Lifecycle stage: None → Staging → Production → Archived

MLflow Model Registry — standard. For enterprise — Vertex AI Model Registry (GCP), SageMaker Model Registry (AWS), Azure ML Model Registry.

Model promotion through stages: automatically move model to Staging after successful eval, then manual or automatic (during A/B test) promotion to Production. Rollback — switch to previous Production version in seconds.

Serving: From FastAPI to Triton Inference Server

Simple case. FastAPI + PyTorch/ONNX on one server — 80% of production ML deployments are exactly that. Sufficient for most tasks with load up to 100 req/s.

from fastapi import FastAPI
import onnxruntime as ort

app = FastAPI()
session = ort.InferenceSession("model.onnx", providers=["CUDAExecutionProvider"])

@app.post("/predict")
async def predict(request: PredictRequest):
    inputs = preprocess(request.text)
    outputs = session.run(None, {"input_ids": inputs})
    return {"label": postprocess(outputs)}

Triton Inference Server — production standard for high loads (500+ req/s). Dynamic batching, concurrent model execution, model ensemble. Supports TensorRT, ONNX, PyTorch TorchScript, TensorFlow SavedModel.

KServe — Kubernetes-native ML serving with autoscaling, canary deployments, A/B testing out of the box. Scale-to-zero for inactive models — savings on infrastructure up to 40% annually for a project with 10 models.

Monitoring: Data Drift, Model Drift, Infrastructure Metrics

Monitoring — what is usually done last and regretted first. Three levels.

Infrastructure monitoring. Latency (P50/P95/P99), throughput (req/s), error rate (4xx, 5xx), GPU/CPU utilization. Prometheus + Grafana — standard. Alert when P99 latency > threshold or error rate > 1%.

Data drift monitoring. Distribution of input data changes over time. Detect via PSI (Population Stability Index) for numerical features: PSI > 0.2 — strong drift. Chi-squared test for categorical, Kolmogorov-Smirnov test for continuous. Evidently AI — open source library with ready-made drift tests.

Model drift monitoring. If ground truth is delayed (e.g., we know conversion after a week) — monitor real metrics. If not — surrogate metrics: distribution of prediction scores, proportion of confident predictions.

Alerting. Three levels: INFO (minor drift, log it), WARNING (significant, notify team), CRITICAL (quality dropped below threshold — automatic switch to fallback model).

Why is data drift monitoring important?

Without it, you learn about model degradation only from user complaints or ringing SLA. A drift alert allows you to retrain the model in advance, before errors start causing losses. In one of our projects, PSI monitoring detected drift 2 days after a data source change — this saved the campaign.

Feature Store

Feature Store solves the training-serving skew problem. If preprocessing during training and inference is implemented in two different places — divergence is inevitable.

A Feature Store is needed when:

• Several models use the same features
• Features are computed from streaming data (real-time)
• Large team with different people on feature engineering and model training

Feast — open source Feature Store. Offline store (S3 + Parquet) for training, online store (Redis, DynamoDB) for low-latency inference. Feature definitions as code, materialization job syncs offline → online.

Tecton (commercial), Vertex AI Feature Store (GCP), SageMaker Feature Store (AWS) — managed options with less ops overhead.

CI/CD for ML

ML CI/CD is regular CI/CD plus specific ML steps.

ML-specific checks in CI:

• Reproducibility check: run training with a fixed seed, result must match
• Data validation: Great Expectations or Pandera on schema/distribution checks
• Model performance check: automatic eval on holdout, block merge if degradation > threshold
• Latency regression test: inference must meet SLA

GitOps for deployment. Merge to main → CI triggers training → eval → if passes → automatic deployment to Staging → smoke tests → manual promotion to Production or automatic upon successful canary.

Tools: GitHub Actions / GitLab CI for CI, ArgoCD for GitOps deployment on Kubernetes.

What's Included in MLOps Platform Development

We provide a full cycle of work, documentation, and team training.

Total time from audit to full MLOps platform: 3–5 months. Also possible phased launch: basic level (tracking + serving) in 4–6 weeks.

Cost is calculated individually based on data volume, number of models, and infrastructure requirements. Order an MLOps infrastructure audit — get a roadmap in 1–2 weeks. Contact us for a project assessment — we will send a preliminary estimate within 2 business days.

Note: warranty on architectural solutions — 12 months. We provide integration certificates with major cloud providers (AWS, GCP, Azure). During our work, we have not lost a single client after the first implementation — the experience of 50+ successful MLOps projects speaks for itself. Get a consultation on building an MLOps platform today.