Search:
Match:
3 results
Research#llm📝 BlogAnalyzed: Dec 25, 2025 17:32

From Social Implementation of AI Agents to Verification of Next-Generation Models: The 25-Day Trajectory of "MEKIKI X AI Hackathon Mogumogu Advent Calendar"

Published:Dec 25, 2025 10:23
1 min read
Zenn AI

Analysis

This article discusses the "MEKIKI X AI Hackathon Mogumogu Advent Calendar," a 25-day initiative focused on AI research and development. It highlights the activities of an AI engineer from NTT Data who initiated the "AI Hackathon/Mokumoku Study Group," starting with an AI hackathon involving Kubernetes GPU clusters on Macs at McDonald's. The project, known as MEKIKI, involves researching and deploying advanced AI technologies. The Advent Calendar involved contributions from members of the study group and external collaborators from NTT Data Advanced Technology and NTT Technocross, showcasing a collaborative effort in exploring AI's potential and practical applications.
Reference

“MEKIKI X AI ハッカソンもぐもぐ勉強会 Advent Calendar 2025 の 25 日目を担当する自称 "NTTデータ3大ミステリーの一つ" とされる葬送のAIエンジニアです。”

PermalinkZenn AI
Research#llm📝 BlogAnalyzed: Dec 24, 2025 22:22

LLM Quantization Day 25: Summary and Future Prospects

Published:Dec 24, 2025 22:08
1 min read
Qiita LLM

Analysis

This article, likely the final installment of a 25-day series on LLM quantization, summarizes the key learnings and explores future trends in the field. Given its placement in an Advent calendar format, it likely provides a high-level overview rather than deep technical dives. The focus on both theory and implementation suggests a practical approach to understanding LLM quantization. The mention of "latest technologies" indicates an awareness of the rapidly evolving landscape of AI model optimization. It would be beneficial to know the specific areas of future prospects that are discussed, such as advancements in quantization techniques, hardware acceleration, or applications in specific domains.
Reference

“LLM quantization from theory to implementation.”

PermalinkQiita LLM
Research#llm📝 BlogAnalyzed: Dec 25, 2025 05:31

Security Analysis LLM Agent in Go (25): Towards Automating Severity Assessment

Published:Dec 24, 2025 21:31
1 min read
Zenn LLM

Analysis

This article concludes a 25-day advent calendar series on building a security analysis LLM agent using Go. It focuses on future plans rather than implementation, specifically addressing the automation of severity assessment for security alerts. The author outlines this as a crucial, yet unrealized, feature of the LLM agent developed throughout the series. The article serves as a roadmap for future development, expressing hope that the author or others will implement this functionality in the coming year. It's a forward-looking piece, highlighting the next steps in enhancing the agent's capabilities.
Reference

“This is a concept that the author is about to work on, and it describes how to further advance the LLM agent implemented in this advent calendar.”

PermalinkZenn LLM