Chromatography data platform · powered by ML

Evo_AI

From a raw .dat to an analytical decision.

Evo_AI turns your chromatographs' binary files into peaks, control charts and anomaly alerts — automatically. Evoelution's analytical intelligence platform, built for pharmaceutical labs.

Built for the pharmaceutical industry
raw_signal.dat · sin procesar
Unfiltered noise
Unresolved peaks
Drifting baseline
0.0 min● no integrado8.0 min

The problem

From raw data to resolved

Proprietary binary|Hard to audit
Every run stays as a .dat binary on the instrument's PC
Proprietary formats and unprocessed signal block any analysis
No traceability or comparison across sequences
Quality control is reactive, never predictive
Why chromatography needs to be replatformed →
evo_ai · procesado
AnálisisTrendingInsights
chromatogram.dat● integrado
RTÁreaFWHMSNR
2.4148,2090.08312:1
3.8731,0040.11204:1
5.1212,8470.0996:1
12%
anomaly score

Built for the rigor of a regulated lab

Waters Empower 3
native .dat parser
< 60 s
from file to report
±2σ / ±3σ
Levey-Jennings limits
0–100%
anomaly score per run

The platform

One pipeline, from raw signal to insight

Every run flows through the same deterministic, auditable path.

01

Ingest

.dat

Drag & drop upload, single or batch. Native Waters Empower 3 parser.

02

Signal processing

SciPy

Savitzky-Golay smoothing and ALS baseline correction (Eilers & Boelens).

03

Peak detection

find_peaks

RT, height, area, FWHM and SNR per peak, with parameters tunable live.

04

Quality control

Levey-Jennings

Control charts with ±2σ / ±3σ limits and configurable specifications.

05

ML Insights

scikit-learn

Unsupervised anomaly detection and automatic run clustering.

Capabilities

The whole analytical cycle, in one platform

Designed with and for analytical chemists.

Single analysis

Interactive chromatogram: raw, smoothed, baseline and corrected signal. Peak table per run.

Trending & QC

Levey-Jennings charts, automatic statistical limits and an alert panel linking straight to each run.

ML Insights

Isolation Forest for anomalies, 2D PCA of the feature space and KMeans to cluster similar runs.

Auditable reports

Export to PDF with Evoelution branding and CSV, ready for quality review.

API-first

Documented FastAPI backend: integrate Evo_AI with your LIMS or existing workflows.

Simple deployment

Docker Compose + Nginx. On-premise or in the cloud — you decide where the data lives.

Intelligence

Machine learning an analyst can understand

No black boxes. Evo_AI uses interpretable, industry-standard methods to flag which run is off-pattern and why.

Isolation Forest

Unsupervised anomaly detection. Runs ranked with a 0–100% score.

2D PCA

Feature-space projection to visually see what clusters and what drifts away.

KMeans

Automatic clustering that identifies distinct methods or conditions.

Compliance

Built for the pharmaceutical industry

The flow is deterministic, reproducible and auditable — aligned with the data-integrity demands of a regulated lab.

  • Full traceability of every run and parameter
  • Auditable PDF/CSV reports
  • Compatible with IQ/OQ/PQ validation
  • Data lives wherever you choose — on-premise or private cloud

Under the hood

Modern, production-grade stack

Next.js 16FastAPIPython 3.11SciPyscikit-learnPlotly.jsTanStack TableReportLabDockerNginx

Ready to modernize your lab?

Tell us what you need and a specialist will reach out in under 2 hours.