Hackathon Portal
AI Tinkerers - New York City
Sign in

AI Tinkerers - New York City Hackathon

Sign in
DevRel Lead - Restate
Build durable AI agent infrastructure. Drive US remote adoption today.
Founding Applied AI Lead - Paxos Health
Build healthcare AI agents in NYC. Apply now.
All Jobs
Final round winners have been announced. View Results
Team

Alinea

Project Concept

No description has been added yet.

Entry

Status: In Progress

Last saved: February 28 at 6:05 PM EST

View Entry

Team Roster

You must be registered for the event to view the team message board.

Serena Wu Team Lead RSVP Approved

teaching assistant at University of Pennsylvania
Serena Wu – Product Design & Multi-Agent Architecture Serena led the product vision and overall concept design for Alinea. She defined the core idea of embedding structured multi-agent deliberation inside an XMTP-powered group chat. Serena designed the agent personas (Skeptic, Creative, Pragmatist, Coordinator), the breakout-room workflow, and the fixed-round convergence mechanism to ensure demo stability. She focused on aligning the system with the hackathon theme of “Group Agents” and ensured that human-agent collaboration was central to the experience. Serena also helped define the integration strategy between XMTP, Convos, and Kilo Gateway. ⸻ Khang Lee – Product Strategy & Interaction Flow Khang contributed to product ideation and user experience strategy. He helped shape how prompts are interpreted inside the main Convos thread and how breakout summaries are presented back to users in a clear, structured format. Khang worked on refining the interaction flow, ensuring that the group setting remained essential rather than decorative. He also assisted in defining how the Coordinator agent structures task splitting, summary aggregation, and final synthesis to make the discussion feel collaborative and intentional. ⸻ Arpit Garg – Technical Lead & System Implementation Arpit served as the technical lead and implemented the backend architecture. He built the XMTP Agent SDK integration in Node.js, enabling agents to operate as participants within Convos group threads. Arpit implemented multi-agent orchestration logic, including breakout thread handling, fixed-round discussions, and final synthesis coordination. He integrated the Kilo Gateway (OpenAI-compatible API) for LLM inference and ensured stable environment configuration with persistent XMTP identity management. Arpit also handled environment setup, agent wallet generation, and system deployment for a live, demoable prototype.
Serena (Luyao) Wu is a Graduate Student in Computer Science and a Teaching Assistant at the University of Pennsylvania in Philadelphia, Pennsylvania. She is also a member of Rewriting the Code. Her educational background includes studies at the University of Pennsylvania, The State University of New York (SUNY), and The London School of Economics and Political Science (LSE) in fields like Computer and Information Technology, Marketing, and MBA Essentials. Serena has six years of experience and is currently open to full-time work, seeking technical co-founders, speaking opportunities, and sponsorships, while offering help with technical architecture.
Technical co-founders, technical architecture, data engineering, AI/ML engineering, backend systems, distributed streaming, cloud infrastructure, speaking opportunities, sponsorships.
Projects include a distributed Kafka streaming pipeline, an AWS S3 data lake with Glue crawlers, and game AI implementations. Technical work features a multithreaded C++ HTTP server using POSIX sockets and the weMeditate mobile app in Swift. Other active implementations include the innocube_full_stack_project using Flask and Docker, plus machine learning pipelines for K-Means and Random Forest in the Data-Portfolio. Open-source work includes contributions to Canonical Snapcraft.

ARPIT GARG RSVP Approved

Machine Learning Engineer at Google
Arpit Garg is currently an ML engineer on the Gemini Enterprise team at Google, focused on improving search quality. As an applied engineer he does literature review, develops proof of concepts and productionizes promising approaches. With all the recent AI advancements he is becoming more interested in e2e applications and e2e problem solving, evolving into a generalist.
Technical co-founders, founding engineers, speaking opportunities, product review, technical architecture, information retrieval, LLM systems, deep reinforcement learning for robotics, recommender systems, distributed training, machine learning models, enterprise search
Created an app to give moltbots a creative outlet. They can collaborate with each other and draw pixel art on a shared canvas -- moltpix.com. Recently also created plexwrap that creates a year-end review like spotify wrap but for Plex.