AI Tinkerers - Hong Kong Online Meetup (June) - AI Dev Tools & MCP Deep Dive (AA TechTalk) [AI Tinkerers - Hong Kong]

AI Tinkerers - Hong Kong Online Meetup (June) - AI Dev Tools & MCP Deep Dive (AA TechTalk)

Jun
10
Tuesday
Tuesday, June 10th, 2025 9PM to 10PM (HKT)
Virtual Meeting

Event Ended

This event has already taken place.

Attendees 15+ registered
The audience includes an Amazon SDE L5 and a Blockchain patent holder, with most common expertise in AI/ML, RAG implementation, and distributed systems.

🤖 Join us for a casual online session focused on AI-assisted dev tools and a deep dive into the Model Context Protocol (MCP)! 🤖

☝️ Register above to participate in an engaging session designed to foster deeper technical exchange and collaborative problem-solving with fellow builders in Hong Kong and the Greater Bay Area.

About AI Tinkerers Hong Kong and Greater Bay Area

AI Tinkerers is a community for individuals building, tinkering, and working with all things AI. It’s designed for practitioners with technical, machine learning, and entrepreneurial backgrounds who are actively building and working with foundation models, such as large language models (LLMs) and generative AI. The purpose is to promote sharing, facilitate learning and collaborations, and support each other in innovating in the rapidly evolving AI space.

Who is this for and What to Expect:

The AI Tinkerers Hong Kong meetups are designed for individuals with technical, machine learning, and entrepreneurial backgrounds who are actively building and working with foundation models, such as large language models (LLMs) and generative AI.

If you’re deeply passionate about utilizing AI, creating new use cases, developing LLM-enabled applications, or have hands-on experience in building such systems, this group is the perfect fit for you. Join to connect with like-minded individuals.

For this online session, we’re focusing on AI-assisted dev tools. We will explore tools like Cursor, Claude Code, and claude-task-master. As a starting point, we will collaboratively try to diagnose Andy Green’s claude-task-master + OpenRouter integration, providing a practical, hands-on discussion for active builders.

So What is MCP Anyway?

The Model Context Protocol (MCP), launched by Anthropic in November 2024, is an open-source protocol that standardizes how AI models connect to external data sources and tools, like a universal adapter for AI. Using a JSON-RPC client-server model, MCP enables applications implementing its client interface to access systems such as GitHub, Google Drive, or databases via reusable MCP servers, eliminating the need for custom integrations. This solves the “N×M integration problem,” making AI workflows scalable and interoperable, with adoption by tools like Theia IDE, Sourcegraph’s Cody, and enterprises like Block.

Anthropic’s Claude Code, a command-line coding tool, exemplifies MCP’s power by implementing and extending the protocol to interact with filesystems, GitHub, and more, enabling context-aware coding assistance. While MCP fosters a growing ecosystem with over 5,000 servers and support from OpenAI and Microsoft, it’s not a standalone solution—it’s middleware requiring robust adoption and security. Developers can leverage MCP’s Python, TypeScript, and other SDKs to build or integrate with servers, streamlining AI-driven tasks across diverse platforms.

How does MCP relate to AI Dev Tools?

Because MCP is getting so popular, a lot of AI-assisted dev tools are adding support for MCP so they can seamlessly connect to a wide range of external data sources and tools without building custom integrations. By implementing the MCP client interface, these tools can leverage existing MCP servers to access systems like GitHub, Google Drive, or databases, enabling context-aware features such as real-time code suggestions, automated repository management, or data-driven insights, all while benefiting from MCP’s standardized, scalable, and secure framework.

(Photo) A screenshot of a Google Meet video conference showing multiple participants and a shared screen displaying the 'Repo Prompt' website. A chat message from one of the participants is also visible at the bottom of the screen. Text: Augustin Chan (Presenting, annotating) Home Features Pricing Docs Repo Prompt Your AI Coding Swiss Army Knife Let Repo Prompt structure your prompts and apply AI changes for you. Download for Mac Windows/Linux/Webkit Join Discord Why Repo Prompt? Zipping your whole repo? Too much noise - send only what's needed LLM underperforming? Cut the noise for sharper answers Better AI coding? Select just the files that matter How It Works Today Augustin Chan Andy Green Alexander Payne Wilko Wielinga Kalvin Ng Stanley Yeung Timothy Chau Alexander Payne tokens of all of those files plus your prompt. Then you know which model you're giving it to, you know, like how many tokens it's going to be and then if it's going to be able to handle it, then I was using this mainly with OT pro and that was very good because it has very good context utilization. So I think all the way up to I think 128k tokens. It's still processes them really well, and we'll actually spit out full files. With. 9:46 PM | tzu-jmru-jqv video conference, website presentation, digital interface | digital (computer screen), implying an indoor setting for participants | multiple people in video feeds, a web browser displaying a product website, a chat window with a message, and video conferencing controls | screenshot, digital capture | Colors: #202124, #FFFFFF, #F4B400, #D93025, #5E35B1 Note: The image is a digital capture of a computer screen, depicting a real-world scenario of a video conference call and a website presentation, which falls under the 'photo' category.

🤖 Connect, Collaborate, and Create 🤖

This online session is your chance to:

  • Dive deep into AI dev tools: Explore and discuss practical applications of tools like Cursor, Claude Code, and claude-task-master.
  • Collaborate on technical challenges: Join in debugging a real-world integration issue with Andy Green’s setup, offering and gaining insights.
  • Network with fellow tinkerers: Connect with other active builders, AI engineers, entrepreneurs, and researchers who share your passion for building the future with AI.

Sponsors, Partners, Volunteers

This event is generously supported by the below partners:

Chapter Founder & Lead Organizer: Timothy

Current/Previous Co-Organizer & Volunteers: Lawrence, Ian Chau, Adam Hok, Iulian, Augustin, William, Anne, Ivan, Yanan

Event will be hosted by: Augustin

RSVP Today!

Space is limited to ensure a focused discussion, so don’t miss out on this incredible opportunity to connect with the Hong Kong AI community. RSVP now.


NOTE: We are adding a “classification” so all our Tinkerers know how technical events are going to be. This way you know exactly what you’re signing up for.

Find out more about our classification system here!

Level 4: Interactive Circuit 🎮

Ready for more?

Check out other posts from this blog.

View all posts