Pioneering research at the intersection of psychology, criminology, and artificial intelligence. We build secure, privacy-first tools for academic research and law enforcement.
The CSI Lab develops cutting-edge, privacy-preserving AI architectures specifically designed for environments where data egress is not an option. We train the next generation of researchers and law enforcement professionals to leverage these powerful tools locally, securely, and ethically.
Building infrastructures that run powerful Large Language Models (LLMs) entirely on local, institutional hardware, ensuring zero data leakage to public clouds.
Upskilling academics, analysts, and officers through hands-on training platforms, making complex AI systems accessible without requiring a computer science degree.
Deploying local AI agents to assist in processing vast amounts of unstructured data, from bodycam footage to complex financial records, accelerating critical investigations.
Advanced applications designed for sensitive incident reporting, realtime guidance, and deep analytics.
Helps officers navigate high-stakes incidents with step-by-step guidance, standardized workflows, and real-time analysis to deliver safe, just, and consistent responses to victims
Instantly understand Washington State law in the field. Powered by an evidence-based AI model that pulls directly from the RCW, we eliminate hallucinations and provide highly accurate answers.
Democratizing access to advanced AI tools. Our platforms are designed to bridge the gap between complex machine learning paradigms and practical, everyday investigative workflows.
A comprehensive technical overview of privacy-first AI models available for institutional use. Compare capabilities, hardware requirements, and use cases for models from Meta, Mistral, Alibaba, and more.
Discover our custom visual interface that links workstations directly to the HPC cluster. Experience seamless job composition, secure file management, and direct chat interactions with offline LLMs—no CLI required.
Locally deployed, privacy-first AI models engineered for highly sensitive, air-gapped environments.
A domain-adapted dense retrieval model fine-tuned with contrastive learning (Multiple Negatives Ranking Loss) on 500K+ Washington State legal query–context pairs, optimized for high-accuracy semantic similarity and ranking across complex statutory data.
Our flagship 1024-dimensional retrieval engine, engineered for enterprise-grade legal AI . By leveraging a parameter-rich architecture, it translates natural language scenarios into exact statutory matches, providing the ultimate high-accuracy backbone for complex Washington State RAG pipelines.
Large-scale, meticulously curated datasets accelerating legal IR and semantic search research.
A large-scale, synthetic dataset designed specifically to advance Legal Information Retrieval. Contains thousands of meticulously paired examples mapping hypothetical scenarios to governing WA State Statutes.
A comprehensive, structured dataset containing the full text of all permanent laws in force in the State of Washington as of 2026. Designed specifically to facilitate legal Natural Language Processing (NLP) research.
Contact the lab to discuss partnership, training opportunities, or deployment strategies for your organization.
