Data Science & Machine Learning

The full technique library

Every discipline behind the six capabilities, across the whole spectrum — classical statistics to deep learning. We pick the one that fits your problem, and prove it on your data.

0+

techniques

0

disciplines

0

capabilities

the whole spectrum

classical statistics → deep learning · not just the fashionable end

106 classical / statistical89 classical ML109 deep learning

/ browse the six

The six capabilities, in depth

Pick a capability — its plain-English job, the questions it answers, and the full toolbox of techniques scroll alongside.

the colour is the tierclassical / statisticalclassical MLdeep learning
01

Predict

what happens next

65techniques · 3 disciplines
22 classical20 ml23 dl

Turn your history into a reliable forward view — demand, risk, failure, churn — so you plan and act before the event, not after it.

answers questions like

  • →“Which customers are about to leave — and who's worth saving?”
  • →“What will demand be next quarter, by SKU and region?”
  • →“Which machine, payment or patient is about to turn critical?”

the toolbox — 65 techniques, classical to deep learning

Supervised & tabular

24 techniques

GLMs (Poisson/Tweedie)Logistic / linearRidge / Lasso / Elastic-NetGAMsCox survivalNaïve BayesDiscriminant analysisCalibration (Platt / isotonic)XGBoostLightGBMCatBoostRandom forestSVMkNNRandom survival forestStacking / blendingTabPFNFT-TransformerSAINTTabNetNODETabRDeepSurvDeepHit

Time-series & forecasting

19 techniques

ARIMA / SARIMAXETS / Holt-WintersState-space / KalmanThetaVAR / VECMGARCH / EGARCHCroston (intermittent)Lag-boosted LightGBMProphetHierarchical reconciliation (MinT)Conformalised intervalsFFORMA ensemblesDeepARTemporal Fusion TransformerN-HiTS / N-BEATSPatchTSTiTransformerTimesNetFoundation: TimesFM / Chronos / Moirai / TimeGPT

Anomaly & outlier detection

22 techniques

z-score / MADEWMA / CUSUMGrubbs / ESDSPC control chartsMahalanobisSeasonal-hybrid ESDMatrix profileIsolation ForestLOFOne-Class SVMElliptic EnvelopeHBOSECOD / COPODkNN outlier (PyOD)Autoencoder / VAEDeep SVDDUSADAnomaly-TransformerTranADDAGMMGANomalyDeepSAD

/ the point

Breadth is the insurance.

A team that only knows deep learning reaches for it even when a GLM would win. We hold the whole library — so your problem gets the technique that actually fits, not the one we happen to like.

Back to the service

production & monitoring: AIOps / MLOps · generative & agents: Generative AI · the data foundation: AI Data Infrastructure