Drift

Dataset QA, Audit & Drift Detection

Detects bias, gaps, duplication, and distribution shifts using baseline clustering snapshots. This is where infrastructure buyers lean in hard.

video

image

text

Production

29.0K runs

Deploy Recipe

Why This Matters

Drift detection is operational monitoring—not a one-time audit. Baseline clusters become the source of truth for data quality over time.

from mixpeek import Mixpeek

client = Mixpeek(api_key="your-api-key")

# Create baseline cluster snapshot
baseline = client.analytics.cluster(
    collection_id="training_data",
    algorithm="hdbscan",
    snapshot_id="baseline_2024_q1"
)

# Detect drift in new data
drift_report = client.analytics.drift_detection(
    collection_id="training_data",
    baseline_snapshot="baseline_2024_q1",
    current_period={
        "start": "2024-10-01",
        "end": "2024-12-31"
    },
    alert_threshold=0.15
)

# Find outliers (potential novelty or data quality issues)
outliers = client.retrievers.execute(
    retriever_id="outlier-search",
    inputs={
        "drift_score_min": 0.8,
        "cluster_id": None  # Unassigned to any cluster
    }
)

Feature Extractors

Image Embedding Text Embedding Video Embedding

Feature Extractors

Image Embedding

Generate visual embeddings for similarity search and clustering

752K runs

Text Embedding

Extract semantic embeddings from documents, transcripts and text content

827K runs

Video Embedding

Generate vector embeddings for video content

610K runs

Retriever Stages

Enrichment Resources

Clustering

Analytics

Documentation

Drift Detection