ICML Night - Instruct.KR X Exa AI Search Technology Meetup in Seoul
- Date
- 2026-07-04
- Location
- Seoul, South Korea
- Host
- Personal
About this event
ICML Night in Seoul brings together two forces that matter right now in applied AI: strong instruction-tuning and alignment communities, and the rapidly changing world of AI search. If you care about how modern systems retrieve, reason over, and act on information in real products, this meetup gives you a focused place to hear ideas, compare approaches, and talk with people building in the space. This is not a generic networking night. It is a technology-first, in-person meetup designed for people who want sharper conversations about AI systems, search infrastructure, autonomy, and what these tools look like beyond demos. About the Event ICML Night - Instruct.KR X Exa AI Search Technology Meetup in Seoul is an in-person gathering for the local and visiting AI community around a shared interest: how advanced search and instruction-based AI systems are being built, used, and evaluated today. The event sits at the intersection of research-minded thinking and practical engineering, making it relevant whether you work on models, products, developer tools, or AI-enabled workflows. The collaboration between Instruct.KR and Exa AI Search Technology signals a meetup with a clear technical center of gravity. Expect discussion grounded in real problems such as retrieval quality, system design, model behavior, agent-like workflows, and the role search plays in making AI applications more reliable and useful. Rather than trying to cover all of AI at once, this meetup has a narrower and more useful focus. That makes it a strong fit for attendees who want depth over noise and who would rather leave with a few high-value conversations and concrete insights than a stack of vague introductions. Because it is hosted in Seoul and timed around ICML energy, the room is likely to include a mix of local builders, globally minded practitioners, and people actively tracking where AI research meets deployment. The in-person format matters here: the best part of a technical meetup is often the follow-up question, the hallway conversation, or the moment when someone explains how they actually solved a problem. What to Expect You should expect a meetup structure centered on technical exchange and community conversation. While the exact flow may vary, the event is well suited for a combination of short talks, product or system perspectives, and open discussion among attendees who are actively thinking about AI search and autonomous systems. Likely themes in the room include: AI search technology and how search changes when the user is a model, an agent, or a developer building AI-native tools Instruction-driven systems and the practical challenges of making model behavior useful, steerable, and dependable Retrieval and context including what it takes to surface the right information at the right time Autonomy and agents especially where search, tools, and reasoning interact in real workflows Product and engineering tradeoffs such as latency, quality, evaluation, and user trust In addition to any formal sessions, expect the in-person discussion to be a major part of the value. Meetups like this work best when attendees bring questions from their own work: what breaks in production, what still feels unsolved, and what people are seeing across different stacks and applications. If you are the kind of person who likes to compare architectures, ask how systems are evaluated, or understand where current tooling helps and where it falls short, this event should give you plenty to engage with. You do not need to come with a polished point of view, but you will get more out of it if you arrive ready to talk specifics. Why Attend This meetup is valuable because it concentrates attention on a part of the AI stack that is becoming increasingly important. Search is no longer just a user-facing product category; it is a core capability inside many AI systems. When models need current information, better grounding, or access to a broader knowledge surface, search becomes part of the architecture. For attendees working on AI products, that makes the topic immediately practical. You can use this event to better understand how others think about retrieval quality, context assembly, tool use, and system reliability. Even a single good conversation can help clarify your own roadmap or expose assumptions in how you are building. For researchers and research-adjacent practitioners, the value is slightly different but equally strong. This is a chance to hear how ideas travel from papers and benchmarks into engineering constraints, product decisions, and end-user behavior. Seeing that translation clearly is useful whether you are designing experiments, evaluating model behavior, or deciding what problems are worth tackling next. There is also simple value in being in the right room. AI moves quickly, but not every event produces meaningful discussion. A focused meetup attracts attendees who care enough about the topic to go beyond headlines. That usually leads to better questions, more useful introductions, and conversations that continue after the event ends. Practical Details This event is in person in South Korea, with Seoul as the focal point for the meetup. If you value spontaneous conversation, direct technical back-and-forth, and the ability to meet collaborators face to face, the format is a major advantage. It takes place on Saturday, July 4 at 12:00 PM GMT+9. A midday Saturday time slot makes it accessible for attendees who want to participate without the pressure of a weekday schedule, and it gives room for the event to function as both a learning session and a networking gathering. A few useful expectations to keep in mind: This is a technology-centered meetup, so you will get the most from it if you are comfortable discussing concrete tools, systems, or ideas In-person attendance matters, especially if your goal is to build relationships with other people in AI, search, and autonomy Come prepared with questions, whether they relate to research, engineering, product strategy, or the limits of current approaches Expect a mixed audience, likely spanning developers, founders, researchers, and technically curious operators If you are in Seoul or can get there easily, this is the kind of event where being physically present will likely create more value than passively following the topic online. The subject matter is timely, the format is direct, and the audience is likely to be full of people thinking seriously about what AI systems need in order to be truly useful.
Who should attend
If your work or curiosity sits anywhere near AI systems, retrieval, search, or autonomy, this meetup is built for you. - **You build AI products or developer tools** and want better intuition for how search, retrieval, and instruction-driven behavior fit into real systems. - **You are an ML engineer, research engineer, or applied researcher** looking to compare practical approaches to model grounding, context handling, and system evaluation. - **You work on agents, copilots, or autonomous workflows** and need to understand how information access shapes reliability, usefulness, and user trust. - **You are a founder or product lead in AI** and want sharper conversations about what is technically possible versus what is robust enough for actual users. - **You are part of the Seoul tech community or visiting for the broader AI moment around ICML** and want to meet serious practitioners in a focused setting. - **You learn best through direct conversation** and prefer asking detailed questions in person instead of consuming generic online content. You do not need to be an expert in every part of the stack. If you care about how AI systems find, use, and act on information, you will likely find people here worth meeting.