Team
Polacky
Project Concept
No description has been added yet.
Entry
Status: Submitted
Last saved: May 09 at 6:04 PM PDT
Team Roster
Message board not available for this team yet.
Artur Wala Team Lead RSVP Approved
ai tinkerer at {model: guide}
Product vision and system design. Authored the Company Brain concept — the spatial canvas as a GTM operating system with three
▎ interaction channels (silent, canvas feedback, explicit). Defined the node taxonomy, connection semantics, and the wiki-backed persistence model. Drove
▎ architectural decisions: MCP as single protocol surface, canvas as state machine, Claude as orchestrator.
I was working in eCommerce tech since 2016, went from intern to leadership role and during that journey created R&D that was responsible for product spinnofs. Built BERT-based multimodal bots when LLM was not a word (2020).
Oxford Alumni.
- voice voice voice
- open source tools
Enterprise Voice AI Architecture. Trying to fight with Sierra / Decagon with open source.
Łukasz Kryczka RSVP Approved
Software Engineer at Dell
MCP server and inline widget. Built the mcp-use (Manufact) MCP server with 12 tools for canvas CRUD and agent lifecycle.
▎ Implemented the inline React Flow widget that renders directly inside Claude's conversation, including bidirectional feedback — node drag events fire
▎ sendFollowUpMessage so Claude reacts to user canvas interactions. Set up the monorepo, dev tooling, and SSE broadcast layer. Used Claude Code (Anthropic)
▎ as the primary development environment.
CEO of BlazeDev. Building AI on smart glasses that trains physical workers on the job. Won ETHGlobal Cannes 2026, 1st at Best Hacking League '24, 2nd AngelHack Warsaw → Global Finals Singapore '24. Co-built AlkOS — an x86_64 OS from scratch in C++23 that runs Doom. Built a space probe at 15. Won the National Philosophy Olympiad at 16. Brokered corporate sponsorships at 19. Master's student at Warsaw University of Technology. SWE at Dell Technologies. Moving to San Francisco.
AI agents on wearable hardware, smart glasses ecosystem (Meta, Apple, Google), B2B SaaS for physical industries, real-time multimodal AI, voice-first interfaces, computer vision for industrial applications, founder community in SF
Building Vison - smart glasses AI that trains physical workers on the job. A mechanic puts on glasses, asks "which hose is this?" and the AI sees what he sees and answers. Does a step wrong, the AI catches it in real-time. We start in automotive service departments. Also co-created AlkOS, an x86_64 operating system that runs Doom.
Piotr Tyrakowski RSVP Approved
Lead developer at DeusMediaGroup
Agent runtime and AI integration. Implemented the autonomous agent runner using Vercel AI SDK and OpenRouter. Built the agent
▎ event system (typed event log with run IDs), concurrent run management with AbortController, and the tool-calling loop that lets agents mutate the canvas
▎ autonomously. Wired up the polling-based status UI in the AgentNode widget component.
Piotr Tyrakowski is a Lead Developer at ChannelCraft, building AI-native products across
video automation (AutoYT), https://autoyt.net/
YouTube analytics (FunnelYT), - https://funnelyt.com/
and consumer iOS (HairBuddy). - https://apps.apple.com/us/app/hairbuddy-prevent-hair-loss
1st place, ETHGlobal Cannes 2026 — NpmGuard + ENS "Most Creative Use" bounty
3rd place, AI Tinkerers x Google DeepMind — multi-agent system that builds websites over a phone call
1st place (Cybersecurity), BEST Hacking League 9th edition
2nd place (Digital Empowerment), HackWarsaw 2024 / AngelHack
TOP 3 (Wellness & Activities), HackYeah 2024
Global Finalist (TOP 30), AngelHack Grand Finals Singapore
Finalist + Audience Award, Szkoła Pionierów PFR VII
1st place, Trading Bot League by XTB
AI, medical imaging, high-performance computing, blockchain, data analysis, software development, health-tech innovations, GPU-accelerated computing
Currently, I am working on AutoYT - autonomous YouTube agent for business.
AutoYT creates professional YouTube videos with one prompt.
autoyt.net
HairBuddy - mobile app that helps man fight with Hairloss.
Michał Zajączkowski RSVP Approved
AI Engineer at Accenture
Frontend engineering. Built the standalone Next.js web app surface with SSE live updates from the MCP server. Worked on React
▎ Flow node styling, canvas layout, and the dark-mode design system (charcoal background, accent colors per node type). Implemented the MiniMap, Controls,
▎ and edge animations.
▎
Michał Zajączkowski is a GenAI Engineer at Accenture, based in Warsaw, Poland. He is also an AI Engineer focusing on Research & Development, where his work involves developing and evaluating generative AI prototypes and machine learning models using Python, Pandas, scikit-learn, and SQL. Michał studied Data Science at Warsaw University of Technology. He is currently looking for technical and business co-founders and is open to introductions via DM.
Partnerships, technical architecture, generative AI prototypes, machine learning models, model evaluation, model fine-tuning, research-driven development, data science, technical co-founders, business co-founders.
Currently serving as a GenAI Engineer at Accenture, where work involves developing and evaluating generative AI prototypes and machine learning models. Key technical activities include utilizing Python, Pandas, and scikit-learn for data analysis, alongside SQL for data management. Focuses on model evaluation and fine-tuning within the GenAI space, leveraging Jupyter for experimentation and research-driven development in an enterprise R&D environment.