Our team secured 3rd place in the largest track at the Consensus Hong Kong Hackathon 2025. Due to a miscommunication, we initially missed our finalist selection, but persisted in seeking feedback and ultimately got the chance to pitch on stage, securing both recognition and potential funding. Our project, Grand Theft Aptos (GTA), is an AI-powered open-world game that integrates the Aptos blockchain for dynamic NPC interactions and an in-game economy. By leveraging AI and blockchain, we created NPCs with unique personalities and behaviors, enhancing player immersion. Key takeaways from the event include never giving up on opportunities, always doing thorough research, and asking instead of assuming to avoid logistical mishaps. Moving forward, we will compete in major ETHGlobal hackathons, continuing to showcase Malaysia’s innovation in Web3. 

Our Team

Our team consists of Qi Jian from Tunku Abdul Rahman University of Management and Technology where he pursues a Data Science degree, Jing Yuan and Marcus are from Taylor’s University with both majoring in Cybersecurity and AI. Despite active participation in hackathons, all three maintained first-class honors. Qi Jian and Jing Yuan have won 14 hackathons, including local and international hackathons, with an undefeated streak.

Marcus occasionally joins them and has contributed to notable wins. We are interns at ALPHV Technologies, where we developed and maintained cutting-edge tech solutions for clients. For this hackathon, we took on distinct roles, leveraging our specializations to build GTA. Jing Yuan developed smart contract code on the Aptos blockchain, integrating native functions. Derek focused on AI components, creating 3D assets in Blender to implement a multi-agentic system. Marcus led the pitching efforts and developed immersive front-end visuals, scraping data from Twitter to enrich our project. 

The Journey

We arrived in Hong Kong a day before the hackathon, eager to maximize networking opportunities through Luma’s side events. That evening, we attended a Web3 gathering in Causeway Bay, where we met founders of AI and blockchain companies, exchanging insights on decentralized AI hosting, blockchain wearables, and the future of Web3. The discussions provided fresh perspectives that would prove invaluable for the competition.

The next morning, we arrived early at the hackathon venue, taking time to explore the adjacent Consensus Hong Kong Conference. Unlike Devcon, which is deeply technical, this event had a stronger business focus, allowing us to engage with industry professionals and to introduce our universities to key players. We attended sponsor workshops with Aptos, XRP (Ripple), and Polkadot before diving into our hackathon project.

With only 24 hours to build it, we worked relentlessly, running on minimal sleep. Meals came from the venue, and budget accommodations kept our focus entirely on the competition. After submitting our project, we took a brief moment to connect with other participants before the pitching session. Initially, we were not selected as finalists, but after seeking feedback from mentors, we learned that we had actually made the list — organizers simply couldn’t locate us in time. This unexpected twist led to our final pitch on stage, where we secured third place in the largest track of the hackathon.

Before heading home, we took a few hours to explore Hong Kong before boarding our flight back to Kuala Lumpur, marking the end of an intense but rewarding experience. 

Our Project

Our project for this hackathon is Grand Theft Aptos. Grand Theft Aptos (GTA) is an open-world game powered by AI-driven NPCs and Aptos blockchain interactions. Inspired by traditional GTA-style games, we wanted to break the limitations of static NPCs by making them dynamic, evolving, and capable of meaningful interactions. Our goal was to create an immersive world where NPCs are not just background characters but unique AI agents with distinct personalities, interests, and real-world behaviors. By integrating blockchain technology, we also introduced real-world economic dynamics, allowing players to interact with NPCs, participate in on-chain events, and shape their own stories.

Our solution combines AI and blockchain to create a truly interactive experience. We use AI-driven NPC generation based on Twitter data, Aptos blockchain for in-game transactions and NFT ownership, and randomized on-chain events for unpredictability. The tech stack includes Next.js for the front end, Blender for 3D environments, GPT-4o & DALL-E for NPC dialogues and assets, and Move-based smart contracts to power the in-game economy. Features like gas-free transactions, AI-driven interactions, and Web2 logins ensure accessibility, while future expansions include cross-game interoperability, player-driven content, and a continuously expanding game world. 

About the Hackathon

The Consensus Hong Kong Hackathon 2025, organized by EasyA, presented an opportunity to compete against top global talent. With a strong hacker community and manageable travel costs, it was an ideal challenge for our team.

With multiple international hackathon victories between us, Jing Yuan and I invited Marcus to join, confident that our combined expertise would allow us to compete at the highest level.

Representing Malaysia, TARUMT, Taylor’s University, and ALPHV Technologies, we competed against elite teams from China, Korea, Japan, the US, Canada, and Russia. The Aptos track alone had 40 submissions, with Tsinghua University securing first place and a Thiel Fellowship recipient with Naver AI taking second. Placing third behind these global powerhouses was a defining achievement, proving that Malaysian talent belongs on the world stage.

Facing teams backed by Microsoft, TUI, and Weibo, alongside founders with millions in funding and years of experience, we held our own. This victory is more than

personal—it’s a statement that Malaysia is ready to lead in Web3 innovation. This is just the beginning.

Here are the links to the winning projects: Aptos Track Winners, Ripple Track Winners, Polkadot Track Winners