Program Structure

Module 1: Drug Development Data Landscape

• Pipeline overview: discovery → preclinical → clinical → approval → lifecycle.
• Data sources: omics, assays, chemistry, EHR/RWD, imaging, PV, manufacturing.
• Common challenges: missing data, bias, batch effects, noise, governance.
• KPIs: time, cost, success probability, decision quality.

Module 2: Analytics & AI Fundamentals (Practical)

• Supervised/unsupervised learning; features/labels; overfitting.
• Model types: regression, trees, clustering, neural networks (high-level).
• Metrics: AUC, precision/recall, calibration, RMSE; error analysis.
• Data prep: cleaning, imbalance, leakage prevention, train/test split.

Module 3: AI in Discovery (Targets to Leads)

• Target ID: multi-omics and network/knowledge graph concepts.
• Virtual screening: docking support + ML scoring (overview).
• QSAR/ADMET: property prediction and early risk flags.
• Generative design: molecule generation concepts + validation needs.

Module 4: Preclinical Analytics

• Assay analytics: normalization, dose-response curves, hit calling concepts.
• Toxicity prediction: in silico screening logic and limitations.
• Biomarkers: feature selection concepts, signal vs noise.
• Study reports: reproducible analysis and QC checklists.

Module 5: Clinical Trials Analytics

• Trial feasibility: site selection, cohort sizing concepts.
• Patient stratification: risk models, subgroup analysis (intro).
• Endpoints and missingness: practical data handling concepts.
• Real-world evidence: confounding awareness and validation mindset.

Module 6: Pharmacovigilance & Medical Monitoring

• Safety data basics: case intake, coding, duplicate detection concepts.
• Signal detection: disproportionality concepts and triage workflows.
• NLP support: literature triage and case narratives (with human review).
• Reporting: documentation and audit readiness basics.

Module 7: Manufacturing, Quality & Supply Chain Analytics

• PAT/SCADA data concepts; batch monitoring and drift detection.
• Anomaly detection for deviations and OOS/OOT signals (intro).
• Quality-by-Design (QbD): CQAs/CPPs and data links (overview).
• Forecasting: demand/inventory optimization use cases.

Module 8: Validation, Compliance & Responsible AI

• Model validation: performance, robustness, explainability needs by use case.
• Bias and generalization: population shift and model drift.
• Privacy/security: handling sensitive health data (high-level).
• Monitoring: dashboards, alerts, change control, documentation.

Final Project

• Pick one stage (discovery/preclinical/clinical/PV/manufacturing).
• Define problem, data, model approach, KPIs, validation plan, risks.
• Deliverables: solution canvas + workflow diagram + KPI dashboard + governance checklist.