Evolus is a technology partner of the AMD Developer Hackathon: ACT III on lablab.ai

The AMD Developer Hackathon: ACT III is the hackathon organized by lablab.ai with AMD for developers, founders, engineers and builders who want to build AI agents and high-performance applications on AMD GPUs in the cloud. It is a hybrid event: the online build runs from October 12 to 18, 2026, and on October 17 and 18 there are on-site days in Rome, Milan and Imperia.

You can take part on your own or in a team of up to 6 people. You sign up on the event page on lablab.ai with your AMD AI Developer Program account: if you don't have one yet, you create it there.

Evolus is a technology partner of the event: we make our platform available to participants and run a dedicated track.

With Evolus, companies build AI agents that do real work: they read documents, answer from company knowledge, run automations and act on the systems the company already uses, through APIs and MCP.

A hackathon is the right place to put this idea to the test. In a few days a team can go from a real process, such as an incoming order or a support request, to an agent that completes it. We want to see what developers build when the basic building blocks are ready and they can focus on the process.

Every team can activate a free Evolus event workspace:

• free, with no card required, for teams of up to 6 people;
• valid until November 15, 2026, 23:59 Italian time;
• agents run on your team's model: any OpenAI-compatible endpoint, including a model on AMD GPUs through Fireworks AI or vLLM on AMD Developer Cloud. Calls to your model don't use Evolus credits;
• 50,000 platform credits per workspace, for document reading, AI Tools and library indexing;
• up to 60 API requests per minute per workspace;
• everything works from code: a REST API for workflows, extractors, AI Tools and libraries, an OpenAI-compatible endpoint for agent chat, and an MCP server.

The quickstart, code samples in curl, Python and TypeScript, and all the details are on the Evolus developer page.

On the lablab.ai page our track is called Evolus Business Agents on AMD. The challenge: build an AI agent that runs a real business process end to end on Evolus.

Pick a process that a real company runs every day and let an agent run it. The agent takes the input (an email, a PDF, a recording, a web form, a chat), reasons over it with an open model served on AMD, and acts: it runs an Evolus workflow, turns documents into data, uses AI Tools, updates another system, or hands the case to a person when it should.

For this track the main agent must use an open model served on AMD infrastructure, through Fireworks AI or your own vLLM endpoint on AMD Developer Cloud, and the project must use at least two Evolus building blocks: for example workflows, document extractors, AI Tools, knowledge libraries, MCP connectors, sub-agents or the chat widget.

Prizes: months of an Evolus plan and credits for the best teams, and a showcase on evolus.ai for the winners. Requirements, judging criteria and the official prize details are on the event page on lablab.ai.

On October 17 and 18 the hackathon also takes place on site, in three venues: Rome (Link Campus University), Milan (SmartCityLab Milano) and Imperia, at IMPERIAWARE.

Imperia matters to us: Evolus is a product of Codedesign, a company based in Taggia, in the province of Imperia. One of the venues is a few kilometers from home, and it is a chance to meet in person the people building AI agents in Liguria and beyond.

As the lablab.ai page states, on-site participation is subject to approval, and travel and accommodation are not covered.

1. Sign up for the hackathon on the event page on lablab.ai with your AMD AI Developer Program account.
2. Read the Evolus developer page: quickstart, how to connect your model and ready-to-run examples.
3. On October 12, at the kickoff, we publish the event link on the hackathon page: open it, sign in with GitHub, Google, Apple or Microsoft and create your team's free workspace.
4. Create your workspace only from the event link: a workspace created earlier is not an event workspace.
5. Connect your model, create an API key and start building.