jongkwan.dev

Posts

10 posts
  1. LLM Agent Memory Poisoning

  2. LLM Jailbreaks and Automated Red Teaming

  3. An Overview of Diffusion Language Models

  4. Output Guardrails, Observability, and the Right to Erasure

  5. Egress Gateway Defense

  6. Defenses at the Agent, RAG, and MCP Layers

  7. Input-Side PII Anonymization and Reversible Mapping

  8. How Personal Data Leaks, and What Data Governance Has to Cover

  9. Toolformer: A Model That Teaches Itself to Use Tools

  10. Transformer Architecture and LLM Serving Optimization