3 Securing AI in Java: Guardrails, Ethical LLMs, and Compliance Tools

As large language models (LLMs) are integrated into enterprise systems, Java teams face emerging risks, including prompt injection, bias, and data governance failures, which can quickly erode user trust and stall production rollouts. Focusing on real threats identified in today’s OWASP LLM risk landscape, this session delivers hands-on approaches for Java developers to harden their AI-powered applications.

The talk details defence-in-depth strategies: automated bias screening, prompt sanitization techniques, and compliant logging workflows, all implemented using LangChain4j and Jakarta EE for scalable, modular LLM orchestration.

You will walk away with concrete, ready-to-apply patterns and code-level practices to build secure, ethical, and regulation-ready LLM apps in modern Java environments.

Bazlur Rahman

A N M Bazlur Rahman is a Software Engineer with over a decade of specialized experience in Java and related technologies. His expertise has been formally recognized through the prestigious title of Java Champion. Beyond his professional commitments, Mr. Rahman is deeply involved in community outreach and education. He is the founder and current moderator of the Java User Group in Bangladesh, where he has organized educational meetups and conferences since 2013. In the past two years, Mr. Rahman has been an active speaker at various international conferences and Java user groups. His talks have often focused on specialized topics such as concurrency and virtual threads. In addition to his community work, Mr. Rahman also serves as an editor for Java Queue at InfoQ and Foojay.io, both of which are leading platforms for technology content. Notably, he has authored five books on the Java programming language in Bengali, three of which have become best sellers in Bangladesh. He is currently writing a new book with O'Reilly titled "Modern Concurrency in Java". https://learning.oreilly.com/library/view/modern-concurrency-in/9781098165406/

Shaaf Syed

Shaaf Syed is a seasoned Java developer, AI enthusiast, and advocate in the developer community. With deep expertise in Quarkus, Jakarta EE, and modern cloud-native Java, he has helped drive innovation in AI-powered applications, particularly focusing on Retrieval-Augmented Generation (RAG), vector databases, and knowledge graph integrations. Shaaf regularly speaks at major conferences and meetups, sharing practical insights on Java, AI, and application modernization. Passionate about open source and developer education, he mentors and writes technical content.