Course Objective & Scope
Copying production databases into test environments or using real customer transcripts directly for LLM evaluation and fine-tuning creates severe breach exposure and KVKK compliance risks. This course covers irreversible anonymisation techniques and statistically faithful synthetic data generation across both tabular/relational datasets and unstructured text.
Participants generate structured and textual synthetic test data using Synthetic Data Vault (SDV), CTGAN and local LLMs, then validate privacy guarantees by running re-identification and membership inference attack simulations.
Who Should Attend?
- Data engineers, data scientists and AI engineers
- Test data management (TDM) and QA automation teams
- Privacy engineers and data governance specialists
- Prerequisite: proficiency in Python (pandas) and basic statistics/database concepts
Key Takeaways
- Analyse direct and quasi-identifiers and apply k-anonymity, l-diversity and t-closeness transformations
- Generate multi-table synthetic test data preserving referential integrity with SDV and generative models
- Create privacy-safe synthetic customer dialogues and documents for LLM evals and fine-tuning
- Bound memorisation risk using differential privacy mechanisms
- Quantify and report synthetic data fidelity, downstream utility and privacy risk metrics
Curriculum
Masking, Pseudonymisation vs True Anonymisation
- Irreversibility and re-identification testing under KVKK and GDPR standards
- Linkage, homogeneity and background-knowledge attacks
Classical Anonymisation and Differential Privacy
- Applying k-anonymity, l-diversity and t-closeness via generalisation, suppression and micro-aggregation
- Differential privacy fundamentals: epsilon budgets and Laplace/Gaussian mechanisms
Tabular and Relational Synthetic Data Generation
- Multi-table synthesis with Synthetic Data Vault (SDV), Gaussian Copula and CTGAN/TVAE
- Preserving foreign-key relationships, business constraints and rare edge cases
Textual Synthetic Data for LLMs and NLP
- Extracting schemas and personas from real tickets to synthesize zero-PII dialogues
- Generating synthetic documents and Q&A pairs of controlled difficulty for RAG and security evals
Validating Utility and Privacy Guarantees
- Measuring fidelity and utility via Kolmogorov-Smirnov, correlation retention and TSTR benchmarks
- Privacy auditing with Distance to Closest Record (DCR), singling-out and membership inference tests
Other Courses in This Category
-
Teaches how to scrub personal data from LLM and agent telemetry in OpenTelemetry, Langfuse and SIEM logs, and engineer retention policies, access controls and KVKK compliance into observability pipelines.
- Live Course
- 5 modules
-
Covers Agent Goal Hijack—the top OWASP Agentic AI risk—showing how attackers divert autonomous planning loops, persistent memory and multi-step execution, and how to defend against it with intent verification.
- Live Course
- 5 modules
-
Covers tool poisoning, rug pulls, tool shadowing and supply chain risks in Model Context Protocol (MCP) servers, third-party agent tools and A2A integrations, and how to build a secure MCP architecture.
- Live Course
- 5 modules
-
Teaches how to prevent remote code execution (RCE) and injection vulnerabilities caused by executing LLM- and agent-generated Python, SQL, shell or template outputs using AST validation and safe output handling.
- Live Course
- 5 modules
-
Teaches how to isolate code-executing, file-processing and browser-using AI agents inside secure sandboxes using Firecracker microVMs, gVisor, Kata Containers, WebAssembly and strict network egress controls.
- Live Course
- 5 modules
-
Hands-on training on direct prompt injection, jailbreak techniques and system prompt leakage, showing developers how to detect and defend against them with input classifiers, instruction hierarchy and tools such as garak.
- Live Course
- 6 modules
Get in touch
