SEA: Search Engines Amsterdam

SEA (Search Engines Amsterdam) is coming 🎉!

The upcoming edition, our final SEA Talk of 2025, is all about Agentic Search and Reasoning, featuring two exciting speakers: Foyzul Karim from industry and Junde Wu from the University of Oxford.

The session will be held in person, and both invited speakers will be presenting remotely. Here is the zoom link.

Speaker #1: Foyzul Karim

Title: Building the Next Generation of AI Tool Discovery: Leveraging Agentic AI and Vector Search for Semantic Matching
Abstract: The rapid proliferation of AI tools demands a more intelligent discovery mechanism than traditional keyword-based search. This talk will introduce CodiesVibe [link], a full-stack search engine built to address this challenge. We will explore technical architecture, focusing on how we leverage modern AI and web technologies to provide a superior user experience. Key topics include the implementation of LangGraph-based semantic search with a Qdrant vector database for multi-vector similarity matching, advanced MongoDB aggregation pipelines for flexible tool categorization, and the overall scalable system design using React, Node.js, and TypeScript. Attendees will gain insights into deploying a performant, enterprise-ready platform that transforms natural human queries into precise AI tool discovery.
Bio: Foyzul Karim is a seasoned industry professional specializing in architecting and deploying scalable, AI-powered applications. With a strong focus on full-stack development and modern cloud infrastructure, his work centers on creating intelligent systems that solve complex data and search challenges. His current project, CodiesVibe, demonstrates his expertise in integrating vector databases, semantic search agents, and robust web frameworks to deliver high-performance user experiences.

Speaker #2: Junde Wu from University of Oxford
Title: Agentic Reasoning: A Streamlined Framework for Enhancing LLM Reasoning with Agentic Tools
Abstract: We introduce Agentic Reasoning, a framework that enhances large language model (LLM) reasoning by integrating external tool-using agents. Agentic Reasoning dynamically leverages web search, code execution, and structured memory to address complex problems requiring deep research. A key innovation in our framework is the Mind-Map agent, which constructs a structured knowledge graph to store reasoning context and track logical relationships, ensuring coherence in long reasoning chains with extensive tool usage. Additionally, we conduct a comprehensive exploration of the Web-Search agent, leading to a highly effective search mechanism that surpasses all prior approaches. When deployed on DeepSeek-R1, our method achieves a new state-of-the-art (SOTA) among public models and delivers performance comparable to OpenAI Deep Research, the leading proprietary model in this domain. Extensive ablation studies validate the optimal selection of agentic tools and confirm the effectiveness of our Mind-Map and Web-Search agents in enhancing LLM reasoning. Code at [link]
Bio: Junde Wu is a PhD researcher at the University of Oxford focused on Agents, AI safety and its application in healthcare. He has published 50+ papers with 3,000+ citations, earning recognitions such as a CVPR Best Paper nomination, AAAI Most Influential Paper, etc. Previously, he led AI initiatives at SenseTime and Baidu. He’s now building Aline [link], to make teams can run multiple agents with transparency and control.

Counter: SEA Talks #295 and #296.