Center for Accelerating Operational Efficiency (CAOE)

We flew too close to the sun... Ronald McInnes, Jonathan Amdahl, Dario D'Addamio, and I just wrapped up DASSH 2025: Security in an AI World, another Devils Invent competition. Unfortunately, our first-place streak has ended, but I'm still proud to share that we secured a fourth place prize. This competition was a learning experience, particularly in technical communication and presentation delivery. From the jump, we aimed to surpass our past hackathon solutions and be more comprehensive than ever. In that regard, I'm proud to say we did succeed—though to a fault. Ultimately, we were unable to explain each facet of our solution in a digestible manner in the allotted 5 minutes. Nevertheless, I am still extremely proud of our solution and the work our team put in, and I truly think this is our best work yet! Our solution, ORCA, is a platform that would enable employees at DHS to make informed decisions efficiently and securely, strengthening national security. For our unique AI solution, we combined two key features: C3AN architecture and a two-layer ontology (a core ontology for DHS as a whole, with domain-specific ontologies for each suborganization). This combination makes any off-the-shelf LLM model supercharged, using a neuro-symbolic, multi-agentic, and structured semantic understanding systems to reduce hallucinations and increase accuracy, while promoting fast scalability and modularity. On the security side, we wanted to find a way for ORCA to do something that AI has never done in government: access information from multiple levels of security clearance all at once. To make this possible, ORCA is built on top of General Dynamics Mission Systems's Trusted Network Environment (TNE) and Trusted Virtual Environment (TVE), United States Department of Defense and Intelligence Community certified multilevel computing solutions that allow users to simultaneously access multiple security levels and operating systems from a single computer. Furthermore, our CI/CD testing of domain-specific applications of ORCA ensures reliability before deployment, using DeepEval for model evaluation and QLORA for fine-tuning and compression. On Saturday morning, we sat down with a Policy Analyst at DHS who told us about the Committee on Foreign Investment in the US struggles to keep up with reviewing all the documentation for cases requiring review. Seeing CFIUS as a perfect application of ORCA, we created a prototype dashboard to see ORCA in action: https://lnkd.in/g9QwA-Ta Congratulations to the winning teams in each prompt, and thank you to Center for Accelerating Operational Efficiency (CAOE), Soft Target Engineering to Neutralize the Threat Reality, and ASU Design Experiences for hosting this event! If you want to see our pitch deck, the project is added to my LinkedIn profile! C3AN: https://lnkd.in/gSyFp9Ff Amit Sheth, Kaushik Roy, et al. QLORA: https://lnkd.in/gvaDsH_b from Cornell University #DASSH #Ontology #DHS #LLM #GenAI #Safety #Security

I recently participated in ?? DASSH (Designing Actionable Solutions for a Secure Homeland), where my team worked on the problem statement- ?? "How can the Department of Homeland Security use AI, specifically LLMs, to improve operations, particularly those that require significant staff time?" We brainstormed solutions, collaborated intensely, and learned so much about the intersection of AI and public service. Working with a fabulous team and addressing a real-world problem truly made us feel that each idea can have a tangible impact. I am grateful for being able to contribute to discussions on such an important topic! Center for Accelerating Operational Efficiency (CAOE)

?? Thrilled to Share! ?? I'm incredibly excited to announce that my team and I secured 3rd place in the DASSH (Designing Actionable Solutions for a Secure Homeland) Competition 2025! ???? Competing against some of the brightest minds in AI and security domain, we tackled complex data challenges, leveraged cutting-edge AI techniques, and developed an innovative solution for enhancing the operations at U.S. Department of Homeland Security (DHS). Our LLM-RAG based interactive solution to improve the communication within internal teams was well appreciated by the respected panel consisting of DHS leadership, Center for Accelerating Operational Efficiency (CAOE) and SENTRY (Soft Target Engineering to Neutralize the Threat Reality) faculties. This hackathon journey was a true test of our skills in problem solving, machine learning, time management and data analytics -and I couldn't be prouder of what we accomplished! ?? A huge shoutout to my amazing teammates (Ting Zan and Yuchen Pan), mentors, and everyone who supported us along the way. This moment adds a feather in my University at Buffalo cap. This experience has reinforced my passion for data-driven insights and AI for better lives, and I can't wait to apply these learning to future challenges! #DASSH2025 #SecurityInAnAIWorld #DataAnalytics #MachineLearning #UniversityAtBuffalo #Competition #ProudMoment ????

Hi LinkedIn folks, I recently participated in a hackathon hosted by the Center for Accelerating Operational Efficiency (CAOE) at Arizona State University and Soft Target Engineering to Neutralize the Threat Reality called??DASSH (Designing Actionable Solutions for a Secure Homeland.) My incredible team worked on this problem statement: ?? "How can the U.S. Department of Homeland Security use AI, specifically LLMs, to improve operations, particularly those that require significant staff time?" Our Problem Statement: ??"Increased information overload during a disaster affects efficient resource allocation and deployment. Manually processing vast amounts of data from affected areas, assistance requests, and resource availability is time-consuming and delays decision-making." Ideation: ?? LLMs can be used to extract key information like damage assessments, situation reports and infrastructure status, available resources, etc., to generate action items based on priority for a FEMA operator. Our Solution: ?? "Enhancing FEMA's Disaster Response with LLMs." The team worked on reducing significant staff operation time during and after disaster response and coordination with several operators. We brainstormed solutions, collaborated intensely, and learned so much about the intersection of AI and public service. Working with a fabulous team and addressing a real-world problem truly made us feel that each idea can have a tangible impact. I am grateful for being able to contribute to discussions on such an important topic! Here's more about the DAASH hackathon: ?? This event emphasizes developing real-world solutions, regardless of the medium or approach. We encourage teams from all disciplines to participate in this unique opportunity to collaborate with peers, academic experts, and industry professionals while tackling pressing challenges in homeland security. Participants will work alongside academic and industry mentors, exploring solutions beyond coding challenges to tackle homeland security issues. For more information, visit: https://lnkd.in/dqUWsjP5 Here's a certificate of recognition for my team's efforts and some cool UI designs of our concept!

Big congrats to the team at the ASU Global Security Initiative who held a wonderful event yesterday to celebrate their 10th Anniversary. It was inspiring to hear Executive Director Nadya Bliss speak on the impacts of research fostered by the Department of Defense including many inventions — from internet technology to duct tape to microwave ovens — that are an indispensable part of everyday life. It was also great to hear from ASU President Dr. Michael Crow who noted that GSI’s real mission is making the world safe for democracy. And I'm very proud of the team in the Center for Accelerating Operational Efficiency (CAOE) who presented our recent work in the poster session. I know we all appreciate the fine work done by the GSI team and we look forward to ten more years!

Great update on current uses of AI! Chris, thanks so much for kicking off our weekend design challenge for Center for Accelerating Operational Efficiency (CAOE) last night! Our students are off to a great start!!

This weekend, I had the opportunity to participate in the DASSH 2025 hackathon, exploring security in an AI world. My team proposed to mitigate AI-enabled cyber attacks through an approach called Defense in Scale, which uses existing LLMs to generate a dynamic honeypot inside DMZs, mirroring the scale of attack. This sends attackers into a recursive loop, which makes them waste their resources and diverts them from accessing critical ports. This experience allowed me to learn and apply foundational concepts in networks and cybersecurity, which I am excited to further explore. Big thank you to the Center for Accelerating Operational Efficiency (CAOE) for organizing this event.

????That's a wrap on DASSH 2025!???? Read a full recap: https://ow.ly/Gn8Q50VaJVF The Designing Actionable Solutions for a Secure Homeland (DASSH) 2025 Student Design Challenge has officially concluded, delivering inspiring and impactful results. Hosted by the Center for Accelerating Operational Efficiency (CAOE), a center within the ASU Global Security Initiative (GSI), in partnership with Soft Target Engineering to Neutralize the Threat Reality (SENTRY) at Northeastern University, this dynamic, hybrid event challenged participants to develop innovative solutions to real-world homeland security problems. Over three days, 160 students from 15 universities, forming 37 teams, collaborated to tackle some of the most pressing Department of Homeland Security (DHS) challenges related to artificial intelligence (AI). Students engaged with complex problem scenarios exploring AI adoption, operational efficiencies within DHS, and AI-enabled threats to critical infrastructure—all while competing for a share of the $50,000 prize pool. The three problem statements: 1. How can we establish a broader comfort with AI to enhance its usage by DHS personnel and in society at large? 2. How can DHS leverage AI, particularly large language models (LLMs), to improve efficiency in its operations? 3. How do we mitigate the potential for AI-enabled attacks on critical infrastructure? Teams collaborated with mentors and industry experts, receiving guidance throughout the event as they developed and honed their solutions. On the final day, students presented their solutions to a distinguished panel of judges, including DHS professionals, academic leaders, and security experts. ??Let’s Celebrate the Winners ?? These student teams secured top honors for each problem statement, earning cash prizes to be split between the members: Problem Statement 1: ?? First Place: Quartz (ASU) ?? Second Place: NSF Seton Hall University (Seton Hall) ?? Third Place: Manual Intelligence (ASU) Problem Statement 2: ?? First Place: KPS (ASU) ?? Second Place: Team BEC (City College of New York & Rutgers) ?? Third Place: UB GenAI Mavericks (University at Buffalo) ?? Tied for Fourth Place: Golden Bulls (Johnson C. Smith University) Extended Warranty (ASU) ?? Honorable Mention: MoneyWell (ASU) Problem Statement 3: ?? First Place: DASSHing Ducks (Stevens Institute of Technology) ?? Second Place: Captain AmerIca (University of the District of Columbia) ?? Third Place: Cyber Psyche (ASU) DASSH would not have been possible without the guidance and expertise of our academic partners, the planning committee and our distinguished mentors and judges, who generously shared their knowledge and provided valuable feedback to student teams. We extend our deepest gratitude to all who participated, supported, and made DASSH 2025 a success. DHS Science and Technology Directorate #DHS_COE_Research #CAOE #AI #DASSH2025 #InnovationInAction #SecurityInAnAIWorld #FutureLeaders #AIandSecurity

Last day of DASSH 2025! ???? The students have dedicated themselves wholeheartedly this weekend, and today, they are bringing their hard work to life through insightful and impactful presentations. Their dedication, creativity, and problem-solving skills have been truly inspiring. We couldn’t be prouder of their commitment to driving innovation and advancing research in homeland security. DHS Science and Technology Directorate #DHS_COE_Research #CAOE #DASSH2025 #InnovationInAction #SecurityInAnAIWorld #FutureLeaders #CollaborateToInnovate

Last day of DASSH 2025! ???? The students have dedicated themselves wholeheartedly this weekend, and today, they are bringing their hard work to life through insightful and impactful presentations. Their dedication, creativity, and problem-solving skills have been truly inspiring. We couldn’t be prouder of their commitment to driving innovation and advancing research in homeland security. DHS Science and Technology Directorate #DHS_COE_Research #CAOE #DASSH2025 #InnovationInAction #SecurityInAnAIWorld #FutureLeaders #CollaborateToInnovate