IICTI Certified Data Scientist

Holders demonstrate that they can apply and evaluate online experimentation and advanced A/B testing, causal inference and observational study design, bayesian modelling and probabilistic programming and advanced statistical learning and model validation, plus 7 further areas, in realistic professional scenarios, as defined in the AQSCERT scheme for Data Scientist.

Who it is for: Professionals working, or preparing to work, in artificial intelligence roles.

1. 01Online experimentation and advanced A/B testing9% of the exam

   • Design and evaluate experiments that yield valid, decision-ready inference under real-world constraints.
   • Scope: Power analysis and MDE, variance reduction (CUPED), sequential and always-valid testing, multiple-testing correction, interference and network effects, switchback and cluster designs, SRM detection and multi-armed bandits.
2. 02Causal inference and observational study design7% of the exam

   • Evaluate causal claims and design identification strategies that defend against confounding and bias.
   • Scope: Potential outcomes and DAGs, identification strategies, propensity scores and doubly robust estimation, difference-in-differences and synthetic control, regression discontinuity, instrumental variables, heterogeneous treatment effects (meta-learners, causal forests).
3. 03Bayesian modelling and probabilistic programming6% of the exam

   • Design Bayesian models that quantify uncertainty and support principled decision-making.
   • Scope: Hierarchical models, prior elicitation, MCMC and variational inference (Stan, PyMC, NumPyro), convergence diagnostics, posterior predictive checks, model comparison (LOO, WAIC) and decision analysis.
4. 04Advanced statistical learning and model validation23% of the exam

   • Evaluate model designs and validation schemes that produce reliable out-of-sample performance estimates.
   • Scope: Gradient-boosted trees, regularisation, bias-variance trade-offs, nested and time-aware cross-validation, leakage detection, probability calibration, imbalanced learning and cost-sensitive evaluation.
5. 05Time-series forecasting and demand modelling6% of the exam

   • Design forecasting systems with calibrated uncertainty and robust backtesting for operational decisions.
   • Scope: State-space and ETS/ARIMA models, hierarchical reconciliation, global ML forecasters (LightGBM, N-BEATS, TFT, foundation models), probabilistic forecasts, backtesting and forecast-value-added analysis.
6. 06Deep learning, representation learning and foundation models8% of the exam

   • Evaluate when and how to apply deep learning and foundation models and design rigorous evaluation for them.
   • Scope: Transformer architectures, embeddings and transfer learning, fine-tuning (LoRA/PEFT), retrieval-augmented generation, LLM evaluation and prompt optimisation, multimodal models and compute/cost trade-offs.
7. 07Feature engineering, data quality and large-scale data processing11% of the exam

   • Design data pipelines and feature sets that are correct, reproducible and free of temporal leakage.
   • Scope: Feature stores, point-in-time correct joins, data contracts and validation (Great Expectations, Pandera), distributed processing (Spark, Polars, DuckDB), sampling and survey weighting, missing-data mechanisms and imputation.
8. 08Model interpretability, fairness and responsible data science8% of the exam

   • Evaluate models for interpretability, fairness and privacy risk and design proportionate mitigations.
   • Scope: SHAP, partial dependence and counterfactual explanations, fairness metrics and trade-offs, bias mitigation, privacy techniques (differential privacy, k-anonymity), EU AI Act and GDPR obligations for profiling.
9. 09Productionising models and MLOps for data scientists8% of the exam

   • Design the path from prototype to monitored production model with clear ownership and risk controls.
   • Scope: Experiment tracking and model registries (MLflow), reproducible environments, batch vs real-time serving, data and concept drift monitoring, shadow and canary deployment, retraining policies and model risk management (SR 11-7).
10. 10Optimisation, decision science and uplift modelling7% of the exam

    • Design decision systems that translate model outputs into optimal, measurable business actions.
    • Scope: Linear and mixed-integer programming, stochastic optimisation, uplift and treatment-policy learning, recommendation and ranking metrics, simulation and what-if analysis linking predictions to business value.
11. 11Leading data science strategy and stakeholder decision-making7% of the exam

    • Lead data science initiatives from problem framing to adopted decision with credible communication of uncertainty.
    • Scope: Problem framing and value estimation, metric design and north-star trees, communicating uncertainty, analytics portfolio prioritisation, peer review standards, team practices and ethics in decision-making.

Question types

Single choice 100%

Delivery

Online (remote proctored), Test centre, On-site (corporate)

Retakes

Up to three attempts within any 12-month period. A 14-day waiting period applies after a first unsuccessful attempt and 30 days after a second. Each retake requires a new booking and fee. Retakes draw a different randomised form from the question bank.

Provisional cut score — psychometric validation is required before live delivery. Blueprint v1.0.

1. 1

   Register

   Create a free account and start the programme.

2. 2

   Practice

   Online practice test (coming soon).

3. 3

   Training

   Optional

   Online course or classroom — or skip it.

4. 4

   Certification

   Independent exam and certification decision.

Training is optional and does not by itself confer certification. Find a training partner