As Low Earth Orbit (LEO) satellite competition intensifies, satellite communication has evolved from a terrestrial supplement into a core infrastructure for AI and cloud services. Emtar Technologies, a Canadian chip design startup founded by Taiwanese-Canadian entrepreneur Alvis Huang, is emerging as a critical player in this shift. Leveraging his background as a Marconi Young Scholar, Huang has led Emtar to develop groundbreaking 6G Non-Terrestrial Network (NTN) chips that have already garnered support from TSMC and the Canadian government.Breakthrough Performance: The "Private Library" ArchitectureAt the recent TSMC North America Technology Symposium, Emtar conducted a Live Demo of its 6G NTN solutions, demonstrating high strategic value to the semiconductor supply chain. Emtar's chipset - comprising high-performance RF Front-End ICs and Intelligent Beamforming chips - acts as the system's "sensory organs and brain."Unlike traditional architectures where transceivers must "queue" to access shared memory (SRAM), Emtar utilizes a disruptive fully embedded memory design paired with proprietary algorithms. This gives each transceiver a "private library" for instantaneous data scheduling, resulting in: 1. 10x faster tracking and position prediction. 2. 2x higher reception sensitivity. 3. Significant power savings (dozens of watts), solving critical heat dissipation issues for satellite equipment.National Recognition and Global ExpansionEmtar's strategic importance is backed by high-level Canadian endorsement. Emtar participated in a Canadian trade mission to South Korea led by Minister Maninder Sidhu, where the company engaged with potential satellite industry partners. Additionally, Emtar was named "Startup of the Year" by Canada's Semiconductor Council (CSC), an organization featuring industry titans like AMD, Intel, and Qualcomm.The Future: Data Centers in SpaceHuang anticipates the LEO market will reach 2 billion users within seven years, driven by "Data Centers in Space." Cloud providers are eyeing orbital AI deployments to mitigate terrestrial geopolitical risks. This shift demands high-efficiency satellite access for everything from autonomous drones to maritime vessels - these are Emtar's primary target market.With products entering mass production by year-end, Emtar is currently engaging with Taiwanese ODMs and space agencies during COMPUTEX Taipei. As products head toward mass production by the end of the year, Emtar plans to launch its Series A funding round. Huang emphasized that he is looking for investors with Silicon Valley experience who can provide top-down strategic resources to help Emtar connect with tier-one global satellite operators. From a Canadian startup to an industry star, Emtar is poised to leave its mark on the 6G space race, blending Taiwanese entrepreneurial resilience with North American technical innovation.

As the global technology industry descends on Taipei for COMPUTEX 2026 this June, a Canadian startup is preparing to challenge the fundamental architecture of modern processing. Irreversible, a Montreal-based firm with deep roots in quantum computing, is unveiling a "physics-first" analog in-memory computing architecture that promises a staggering 1,000x reduction in power consumption compared to conventional digital processors.Unlike many silicon startups that originate in traditional chip design, Irreversible’s journey began in the highly constrained world of quantum physics. The core team originally specialized in quantum computing, where they were forced to solve computing problems within the extreme limitations of a dilution refrigerator - an environment where even the slightest heat or noise can destroy a fragile quantum state. Co-Founder Dominic Marchand explains that this background led them to become a "computing company that found its way to designing chips," rather than the reverse. This unique DNA pushed the team to strip away decades of architectural abstractions and return to the most basic laws of physics to find the most energy-efficient ways to process information.The current industry obsession with massive Large Language Models (LLMs) has created a significant energy crisis, particularly at the "extreme edge" where devices must operate on microwatt-class power. Irreversible addresses this by sidestepping the Von Neumann bottleneck, the energy-intensive movement of data back and forth between memory and the processor. By performing calculations directly in memory and maintaining a fully analog signal path, Irreversible also eliminates the power-hungry digital-to-analog conversions that often limit hybrid AI chips. This approach recognizes that while digital logic offers noise protection, the energy required to strictly maintain ones and zeros is a luxury that edge sensors can no longer afford.A critical point of differentiation is how Irreversible compares to other innovators in this space. Marchand notes that while he is proud of the Canadian leadership in analog in-memory compute, Irreversible maintains several distinct advantages. First, the company is memory-agnostic, meaning they are not tied to a single proprietary memory technology and can instead utilize various non-volatile memories and emerging RRAM roadmaps. Second, the company places extraordinary emphasis on its software and simulation tools, which allow their hardware and software teams to work in lockstep. Their proprietary hardware-aware training ensures that neural networks remain accurate by accounting for the inherent variability of analog circuits during the initial training phase.For their up-coming visit to Taipei, Irreversible has set clear strategic objectives to integrate with the world's leading semiconductor ecosystem. A primary goal is establishing high-level connections with semiconductor foundries to gain privileged access to specific memory cells, which are essential for their "physics-first" custom designs. Additionally, the company is actively seeking partnerships with OEMs and solution integrators. By bringing intelligence directly to the sensor site, Irreversible aims to enable "previously impossible" use cases, such as deploying sophisticated AI on small drones or always-on wearable devices that cannot support a traditional GPU. Ultimately, Irreversible arrives in Taipei not just to showcase a chip, but to advocate for a shift in how the world thinks about intelligence. By trading the rigid certainty of digital bits for the natural efficiency of physics, they are proving that the future of AI isn't just about more power - it's about more efficient computing.

In the high-stakes world of global security and emergency response, the shift toward "video-centric" operations has created a massive technical bottleneck: the struggle to transmit high-quality data over narrow, unreliable bandwidth. Secure City Solutions, a Canadian fast-growing company, is bridging this gap between military-grade demands and smart city infrastructure.In an exclusive interview, Siva Kumar, CEO of Secure City Solutions, explained why Taiwan plays an essential role in the company's global expansion, and why he has signed up to attend COMPUTEX 2026 in Taipei in June. "We not only want to address the Taiwan market, but we're also looking for hardware manufacturers for our global deployment."The company was born from a specific challenge faced by founders with deep military and defense backgrounds, including former General Dynamics leadership and a Colonel in the Canadian defense forces. They recognized that whether in a military conflict or a law enforcement pursuit, personnel often struggled to send big data over radios with low bandwidth. This led to the development of a unique solution designed to deliver forensic-quality video from one point to another without losing the essential details required for legal and operational use.At the heart of their offering is the Omni Compressor, a neural-type algorithm that drastically shrinks the digital footprint of video data. While traditional compression often drops frames or reduces resolution to save space, Secure City's technology maintains the original frame rate and resolution. This is a critical distinction for law enforcement, as compromised video quality is often inadmissible in court. Beyond the legal sector, the compression allows commercial entities like banks to store eight to nine times more footage on existing hardware without losing clarity. The company also claims to reduce costs by 75% compared to other solutions.The real-world impact of this technology is already visible in major global deployments, such as with the Dubai Police and over 45 other law enforcement agencies. The software allows police units to share live video from patrol cars or body cameras over weak wireless spectrums, ensuring that backup units can monitor officers entering dangerous areas. Firefighters have also adopted the technology, using helmet-mounted cameras to transmit live feeds to commanders who guide them through burning structures to rescue civilians. Even in rural areas where 5G is unavailable, the algorithm automatically adjusts to available bandwidth and uses high error correction to keep feeds stable despite network noise or jitter.As artificial intelligence becomes more prevalent in surveillance, Secure City Solutions serves as a vital performance booster. By reducing data sizes - for instance, from 100MB to 10MB - while maintaining original quality, the software allows AI models to process information and produce results much faster than they could with uncompressed files.Looking toward the future, Kumar is exploring strategic partnerships in Taiwan to address local needs for data sovereignty and hardware manufacturing. While the company has been successfully bootstrapped by its conservative, veteran leadership, they are now open to strategic investors to fuel a more rapid global expansion into new verticals like transportation and medical services. Secure City Solutions aims to ensure that no matter how narrow the pipe, the most critical data always gets through.

Montreal-based AON3D is setting a new standard through its mastery of high-performance materials and precision 3D printing technology.Co-founded in 2015 by Andrew Walker, Randeep Singh, and Kevin Han - who started the company in his family's basement - AON3D has evolved into a global leader in high-performance additive manufacturing.With an eye on the Taiwan market at COMPUTEX 2026, Han and his team are ready to bridge the gap between complex aerospace technology and the agile SME ecosystem.The Materials Engineer's VisionKevin Han's journey began at McGill University with a background in materials engineering. After operating as a service bureau, Han recognized a gap in the market for machines capable of handling specialized materials. Through multiple product iterations, AON3D today offers it's Hylo High-Temperature 3D Printer, along with Basis, it's advanced physics simulation software for additive manufacturing. Hylo and Basis: AI-Infused and Physics-BasedAON3D's product suite offers an AI-infused manufacturing solution that reduces the trial and error usually experienced in additive manufacturing processes. "What we do is actually model out at the physics level what's going to happen as you run the print job," says Han. "Our technology creates a digital twin of the print, meaning we can use simulation to identify process irregularities that lead to hidden defects, instead of in post-production."Within the Basis platform, simulated data and real data are also compared to offer automatic optimizations. The Power of "Open Materials"AON3D's primary competitive advantage is its "Open Materials" philosophy. Unlike competitors that "lock" users into proprietary, expensive filament spools - much like the cartridges on a paper-based printer - AON3D's platform is supply-agnostic. "We support the full gamut of industrial polymers, but many customers are most interested by high-performance varieties like PEEK, PEKK, and PEI (Ultem)," Han explains. "This includes their carbon and glass-fibre variants, where strength and lightweighting benefits most appeal to demanding industries like aerospace and defense." From NASA to the Factory FloorAON3D's credentials extend to outer space. The company has worked with the Canadian Space Agency (CSA) to print components for the International Space Station, and their parts were aboard the Artemis 1 mission.Closer to home, AON3D's solutions are used by customers like Boeing, Lockheed Martin, Northrup Grumman, and more in aerospace, while also offering benefits to automotive, energy, and general manufacturing. One automotive customer saw full payback under 2 months for their first Hylo purchase, and is eagerly awaiting more. Leveraging Taiwan's EcosystemAt COMPUTEX 2026, AON3D aims to connect with Taiwan's semiconductor packaging and testing sectors. Beyond chips, they see massive potential in Taiwan's drone industry and medical prosthesis field. Hylo's ability to "light-weight" components makes it ideal for rapid drone iteration.  "We want to bring capabilities to a group that didn't have them before," Han concludes. AON3D isn't just selling a printer; they are offering a gateway to the next generation of industrial manufacturing.

Test Research, Inc. (TRI), the leading provider of Test and Inspection solutions for the electronics manufacturing industry, is proud to announce the launch of the TR7950Q SII Series. This highly modular platform is a dedicated solution for Back End Process and Advanced Packaging Inspection, ranging from patterning to wafer saw, and is engineered to set new benchmarks in wafer inspection and micro-measurement metrology.The AI-powered Wafer Metrology and Inspection Platform, TR7950Q SII, is built on a high-stability granite platform and the system supports 6" to 12" wafers. The platform features robust Automated Visual Inspection (AVI) for high-speed detection of surface defects, including particles, scratches, chipping, contamination, and foreign materials.The optional Short-Wave Infrared (SWIR) module allows the system to penetrate silicon to detect hidden inner cracks and subsurface defects invisible to standard sensors. For high-detail requirements, the platform offers 0.5 µm or 1 µm high-resolution imaging via the 3D DFF (Depth from Focus) module.The TR7950Q SII provides high-precision metrology for wafer thickness, top-side warpage, and complex surface topography, alongside high-speed sensing for Through-Silicon Via (TSV) depth, trench dimensions, thin film, and Chiplet metrology. Please visit the link to learn more about the TR7950Q SII.Credit: TRI

Global Unichip Corp. (GUC), the Advanced ASIC Leader, today announced a strategic technical collaboration with Wiwynn, an innovative cloud IT infrastructure provider for data centers. This collaboration integrates GUC's flagship SoC design and 2.5D/3D advanced packaging with Wiwynn's expertise in rack-scale system integration, liquid cooling and optical interconnect. Together, the collaboration enables hyperscale customers to transition more efficiently from silicon definition to deployment-ready AI infrastructure.AI clusters continue to scale in performance, bandwidth and power density, hyperscalers must increasingly evaluate silicon, packaging, interconnect, thermal and rack-level design choices much earlier in the development cycle. Through this collaboration, GUC and Wiwynn are aligning key technology pillars, including leading-edge ASIC implementation, 2.5D/3D advanced packaging, optical I/O, power delivery, thermal architecture, manufacturability, serviceability and rack-scale integration. By addressing these factors holistically at the outset, the collaboration partners aim to reduce integration complexity, improve development efficiency and accelerate the transition from silicon-ready innovation to system-ready AI infrastructure."As AI infrastructure evolves beyond chip-level optimization and scale-up networks push the limits of conventional electrical interconnects, close alignment across silicon to system architecture become critical," said Aditya Raina, Chief of Marketing of GUC. "By collaborating with Wiwynn, we are helping hyperscale customers evaluate critical system-level tradeoffs earlier, integrating optical I/O to deliver the bandwidth and power efficiency required for next generation AI systems. This partnership establishes a more practical, holistic path from flagship ASIC development to deployable, rack-scale AI infrastructure."With deep expertise across board-level innovation, rack-scale integration, and manufacturing, Wiwynn effectively bridges semiconductor innovation with data center deployment," said Tony Wen, Vice President at Wiwynn. "Together with GUC, we are enabling a comprehensive silicon-tosystem approach that delivers scalable, efficient and serviceable AI infrastructure tailored for nextgeneration hyperscale environments."For more information, please visit.

Cincoze has unveiled its all-new CV-200 Series slim-bezel industrial displays. The CV-200 Series features a minimalist profile, a narrow bezel, and industrial-grade reliability for industrial panel PCs and industrial touch monitors. Specifically engineered for modern factory HMI and process visualization, they carefully balance the durability required for harsh environments with seamless equipment integration and intuitive operation. The modular design of the CV-200 Series offers screen sizes from 10 to 21.5 inches for over 40 possible configurations. The first release is the 21.5" Full HD models with almost ten configuration options for various application needs.Ultra-Slim Bezel, High Visibility, and an Intuitive User ExperienceThe CV-200 Series offers clear visuals and smooth operation, and integrates easily into production line equipment. Its slim, die-cast aluminum alloy frame has a bezel less than 3mm wide, increasing the display area without changing existing equipment setup. The Full HD screen and 178° wide viewing angle ensure clear and crisp readability from any position. Every model features a projective capacitive (P-Cap) touchscreen with an anti-glare (AG) coating for the clearest images, even in high-brightness indoor lighting conditions. Touch response is fast and precise, making daily HMI operation smoother and more natural.Rugged and Durable for Industrial and Humid EnvironmentsThe CV-200 Series is built to handle harsh, humid industrial environments. It has an IP66-rated front panel and Wet Tracking technology, so the touchscreen works reliably even with wet fingers or splashes of water. The backlight lasts up to 50,000 hours, and a 7H hardness Glass-Glass (GG) panel adds durability. The CV-200 Series meets the IEC 61000-6-4 industrial EMC standard, ensuring stable, long-term operation and giving operators total peace of mind.Flexible Modular DesignCincoze's exclusive Convertible Display System (CDS) technology lets you pair the CV-200 Series with embedded computer modules (P2000/P1000 Series) or monitor modules (M1000 Series). Customers can configure the system as either an industrial panel PC or an industrial touch monitor, depending on display size, computing performance, and functional requirements. This plug-and-play design simplifies deployment and maintenance. If repairs are needed, only a single module needs to be replaced, cutting downtime, lowering maintenance costs, and streamlining future upgrades.

Chroma ATE Inc. successfully concluded the 3rd Chroma Paper Award on March 19, marking another milestone in the company's ongoing commitment to industry-academia collaboration. Organized in partnership with National Taiwan University of Science and Technology (Taiwan Tech), the Chroma Paper Award provides an exchange platform that accelerates the translation of research into real-world applications, fostering innovation and the cultivation of key talent.As high-performance computing (HPC) and AI applications continue to advance rapidly, demand for test and measurement technologies is rising across AI chips, HPC, and data centers. In this era of lightning-fast iteration, testing and validation have become critical enablers of system performance and reliability. With long-standing expertise in this field, Chroma has built comprehensive AI-related testing capabilities spanning the four core pillars of AI infrastructure: compute and data processing, cooling and thermal management, high-speed communications and data transmission, and power and energy management. Through its test and validation solutions, Chroma helps ensure the performance, stability, and reliability of AI systems, playing a vital enabling role across the AI value chain.Aligned with Chroma's core technology development priorities, this year's competition featured two main categories: Power Electronics-Related Technologies and Semiconductor Testing-Related Technologies. Each category included a Top Prize (NT$200,000), Excellence Prize (NT$100,000), and Merit Prize (NT$20,000), with total prize funding reaching nearly NT$1.5 million. A total of 101 papers were submitted. Following a multi-stage evaluation process comprising preliminary, secondary, and final reviews, the judging committee selected outstanding research projects distinguished by both technical innovation and strong potential for industrial application.In the Power Electronics-Related Technologies category, the Top Prize was awarded to a team from Taiwan Tech for the project "Development of a High Power Density Power Module for AI Server Applications." Supervised by Associate Professor Yu-Chen Liu and completed by student Yu-Jun Li, the project directly addresses the growing need for high-efficiency, high-power-density power systems in AI data centers.In the Semiconductor Testing-Related Technologies category, the Top Prize went to a team from National Taiwan University for the project "A Novel Full-Chip Afterpulsing Evaluation Technique for 116 x 160 Ge-on-Si SPAD Array." Supervised by Professor Chao-Hsin Wu and jointly completed by students Jia-Zhen Cai, Ren-Hong Zhang, and Qi'en Chen, the research presents innovative achievements in advanced sensing and semiconductor testing technologies.Following the award ceremony, Chroma invited finalist teams into its R&D labs and testing facilities to see firsthand how research outcomes are transformed into real systems and industrial applications. The visit further strengthened ties between academia and industry.As the AI era continues to evolve at speed, test and measurement serves not only as a gatekeeper of quality but also as a key engine of innovation. Looking ahead, Chroma will continue to deepen its engagement with the global academic community through the Chroma Paper Award and a range of industry-academia collaboration initiatives, building an open and impactful platform that advances key technologies and sustains innovation momentum across the industry.For the full list of award winners, please visit the official Chroma Paper Award website.President Jia-Yush Yen of National Taiwan University of Science and Technology delivers remarks. Credit: ChromaCredit: ChromaFinalist students pose for a group photo with Chroma Foundation Chairman Paul Ying. Credit: ChromaChroma ATE Chairman Leo Huang (left) poses with the Top Prize-winning student Yu-Jun Li. Credit: ChromaChroma ATE Chairman Leo Huang (left) poses with the Top Prize-winning student Jia-Zhen Cai. Credit: ChromaFinalist teams visit Chroma headquarters. Credit: ChromaGroup photo at the 3rd Chroma Paper Award ceremony and banquet. Credit: ChromaChroma ATE Chairman Leo Huang delivers remarks. Credit: ChromaChroma Foundation Chairman Paul Ying delivers remarks. Credit: Chroma

The rapid advancement of generative artificial intelligence (GenAI) has significantly enhanced the efficiency of content creation and dissemination. At the same time, it has accelerated the proliferation of misinformation, manipulated content, and digital fraud, posing increasing challenges to democratic governance, social stability, and the integrity of digital trust and information ecosystems. In this context, achieving a balance between technological innovation and risk governance, while strengthening a trusted information environment, has become a key priority for both government and industry in Taiwan.Concurrently, the global expansion of virtual asset investment has prompted jurisdictions worldwide to strengthen regulatory frameworks governing digital asset transactions. Despite these efforts, fraudulent activities involving virtual assets continue to evolve in both scale and complexity. Malicious actors frequently exploit nominee accounts to conduct layered money laundering schemes, resulting in delayed fraud detection and challenges in asset recovery. Conventional approaches -including static blacklists and rule-based controls-are increasingly insufficient to address these evolving threats. As a result, the development of highly interpretable, AI-enabled early warning mechanisms to support timely risk identification and mitigation has emerged as a critical focus for both policymakers and industry.Against this backdrop, Team Bibi Lab, comprising students from National Tsing Hua University (NTHU), was awarded top honors in the "Virtual Asset Transaction Security" category-sponsored by BitoPro, a cryptocurrency exchange under BitoGroup-at the "Agent for Truth: Disinformation Defense Hackathon." Their project, "AI Anti-Fraud Guardian for Virtual Currency Transactions," demonstrated a high degree of technical innovation and real-world applicability, earning strong recognition from the judging panel.Team Bibi Lab extracted 140 features across 18 categories from KYC, fiat currency, and cryptocurrency transaction data provided by BitoPro. At the core of their approach is an innovative "three-layer risk tracking mechanism," which analyzes shared wallets, shared IP addresses, and indirectly connected users to trace how risk propagates through the transaction network-forming one of the system's most critical signal sources.To ensure high interpretability and support compliance requirements, the team deliberately avoided complex ensemble models that could obscure decision logic. Instead, they adopted LightGBM combined with Focal Loss, enabling the model to focus on hard-to-detect minority cases and effectively address severe data imbalance.In addition, the team constructed a network graph of internal transfers, shared wallets, and common IP addresses using NetworkX. By applying algorithms such as PageRank and community detection, they identified critical relay nodes within fund flows, further strengthening the system's ability to uncover hidden fraud patterns.The competition was powered by Amazon Web Services (AWS), whose cloud infrastructure enabled Team Bibi Lab to build a comprehensive AI-driven anti-fraud system. Their solution features a four-layer interpretability architecture: "Continuous Risk Scoring and Tiering," which classifies users into four risk levels from low to extremely high; "Feature Deviation Analysis," which visualizes how user behavior diverges from baseline norms; "Rule-Based Interpretation," which generates automated textual explanations; and an "AI Risk Diagnosis Report," which leverages the Claude 3.5 Haiku on Amazon Bedrock to generate professional compliance analysis reports.The system is built on a robust AWS cloud stack. The solution is deployed on Amazon Elastic Compute Cloud (Amazon EC2), with raw data and feature matrices stored in Amazon Simple Storage Service (Amazon S3), and AWS Glue handling ETL and feature engineering. AWS Lambda supports batch risk scoring, while AWS Step Functions orchestrates the end-to-end workflow. For high-risk cases, Amazon Simple Notification Service (Amazon SNS) sends real-time alerts to compliance teams, and Amazon CloudWatch ensures system monitoring and alerting-demonstrating the breadth and scalability of AWS cloud services in supporting advanced AI application development.Team Bibi Lab noted that, despite their background in AI development, they had no prior experience with AWS cloud services. Prior to the competition, AWS organized a series of technical workshops-covering AI applications and enterprise data-enabling the team to quickly build proficiency with relevant tools and services.With guidance from BitoPro mentors and industry experts, the team deepened its understanding of cryptocurrency operations. They also incorporated key design principles-including "auditability for compliance personnel" and "automated risk alerting" -into their solution, ultimately delivering a practical, deployable anti-fraud system that contributes to a more resilient and trustworthy digital financial ecosystem.

In Taiwan, scams have evolved from isolated tactics into cross-channel, multi-step attack schemes. Most victims are aware of fraud; rather, they are often driven to make poor decisions under conditions of high pressure, limited information, and tight time constraints. Enabling individuals to access reliable risk assessments and timely guidance before taking critical actions has therefore become an urgent priority.At the same time, rapid advances in generative AI have accelerated the proliferation of sophisticated misinformation, deepfakes, and fraud schemes, posing growing threats to social trust and public safety. Strengthening information verification and risk mitigation through AI has thus emerged as a pressing global challenge.In the "Agent for Truth: Disinformation Defense Hackathon," Team (1), formed by members from the Department of Electrical Engineering at National Taiwan University, was awarded the "Excellent Award" in the "Fraud Identification and Prevention" category, sponsored by Gogolook, for its solution "FakeOff."Fraud tactics continue to evolve rapidly, often leveraging current events to lower public vigilance. For instance, during the tax filing season, deceptive SMS messages tend to surge. Traditional anti-fraud models, which rely heavily on historical datasets, often lack the ability to proactively detect newly emerging fraud patterns.Team (1) explained that "FakeOff" addresses this limitation through a continuous learning and data alignment mechanism. By leveraging web scraping technologies to monitor major news platforms, the system captures real-time events and applies AI to identify content that could potentially be exploited by malicious actors. This approach enables model training and deployment prior to the large-scale spread of fraudulent messages, thereby strengthening early-stage fraud prevention.The competition was powered by Amazon Web Services (AWS), providing cloud infrastructure that enabled teams to build and deploy advanced AI solutions at scale. Team (1) utilized AWS infrastructure to build a comprehensive AI-driven anti-fraud system. The solution was deployed on Amazon Elastic Compute Cloud (Amazon EC2), with large-scale fraud datasets and training data stored in Amazon Simple Storage Service (Amazon S3). It also leveraged Amazon Titan Text Embeddings V2 on Amazon Bedrock for efficient text vectorization, highlighting the breadth and scalability of AWS cloud services in supporting advanced AI application development.From a technical architecture perspective, "FakeOff" integrates the visual language model Claude Sonnet 4.6 to analyze screenshots and textual content. The system adopts a multi-agent collaborative framework built on Claude Haiku 4.5, in which a Function Calling Agent dynamically invokes fraud detection tools, blacklists, and number lookup APIs based on visual cues. A Conclusion Agent then consolidates these outputs to generate interpretable assessment reports and actionable prevention recommendations.To address the core challenge of fraudulent message detection, "FakeOff" incorporates a cross-model alignment and voting mechanism. By integrating leading large language models-including GPT-4o, Claude Sonnet 4.5, Llama 3 70B, Mistral, and Qwen3-30B-the system performs cross-validation to identify models that best capture the nuances of the Chinese language. It further enhances performance by localizing and training on global datasets, effectively addressing the limited availability of fraud-related data in Taiwan.The solution also adopts a continuous learning architecture, combining Amazon Titan Text Embeddings V2 for text vectorization with a neural network classifier. This design enables ongoing model refinement through real-world data and human feedback, ensuring sustained accuracy and adaptability in an evolving fraud landscape.To counter emerging fraud tactics, "FakeOff" autonomously scans recent news across multiple platforms to extract high-risk keywords, enabling early detection of new scam patterns without requiring full model retraining. This forward-looking technical approach was a key factor in earning strong recognition from the judging panel.Team (1) noted that, with on-site support from AWS Solution Architects and mentorship from Gogolook, the competition enabled the team to expand its perspective beyond purely technical development to encompass real-world application scenarios. The team also expressed its aspiration to contribute to fraud prevention and strengthen information security.