Project

Autonome — no-code AI-agent deployment platform

Product engineer

AltLayer's no-code AI-agent deployment platform. Users pick a framework — Eliza, AgentKit-TS, Based Agent, Perplexica, or their own Docker image — configure persona / API keys / env vars, and get a chat-ready agent in roughly five minutes. Supports verifiable execution via TEE (Automata). Lives at apps.autono.meme inside the unified AltLayer Wizard dashboard.

Owned a meaningful share of the Autonome surface inside the Wizard frontend — the framework-discovery gallery, the new-agent deployment flow, the agent metadata editor, the environment-variables dialog, and the Eliza framework's UI integration end.

• Shipped the framework-discovery gallery (AgentFrameworksSection) — a multi-row Swiper carousel that adapts to viewport and template count, with dynamic featured-template assignment so new frameworks like AgentKit-TS can be promoted without component changes.
• Wired up the Eliza framework end-to-end in the UI — character-schema capture, settings-form validation, and a BNB-flavoured template with a custom Yup validator for the agent's web3 private key, catching invalid inputs before they ever reach the runtime.
• Contributed across the new-agent deployment flow and the per-agent metadata editor so non-technical operators can deploy and manage agents without touching Docker, Kubernetes, or YAML.
• Improved the environment-variables dialog with select-field support, so framework authors can declare env vars that pick from a fixed set rather than free-form text — cutting whole classes of misconfiguration.
• Tightened the AaaS shell with floating Telegram and FAQ buttons so users always have a one-click path to community support and in-context help while configuring their first agent.
• Reinforced cross-cutting design-system primitives (FormSelectField clearable prop + padding parity, DateRangePicker granularity options) that Autonome and the rest of the Wizard surface both consume.