Лид AI-направления (Head of AI) at TTM Dev KZ — NeverHard

About Phase: Phase is our own L1 blockchain, built in large part for the world of AI agents. The network is designed for mobile devices: phones can become full nodes. We have the blockchain protocol, mobile apps, and products on top of the network. A separate direction is t ooling and skills for AI agents that let them interact with the Phase blockchain and use what it offers. We want Phase to be convenient blockchain infrastructure for agents: the actions they need available through ready-made skills, and those skills present in the major agent ecosystems and marketplaces. The team is small and senior, working from an office in Almaty. Our principle: agents write the code; people think, set the tasks, and own the outcome. Why this role exists: AI and the agent ecosystem change every week. We want to be the team that adapts first, not the one catching up. Today the CTO covers AI alongside everything else. We need someone for whom this is the core ownership: growing AI inside Phase and building the product for external AI agents. What you'll do: 1. AI vision and strategy Track how models, agents, MCP, skills, marketplaces, and agent frameworks evolve, and turn that into concrete decisions: what we adopt, what we test, and where we go next. 2. Skills for AI agents One of the main areas of ownership. Decide which skills agents need to work with Phase, and take them from idea to working product. You will: design and build skills for agents to interact with the Phase blockchain; brief coding agents and verify the result; improve quality, prompts, and agent workflows; maintain and grow the existing skills. 3. Skill promotion and distribution Not just building a skill, but getting people to actually use it. You'll own: publishing skills to the relevant agent marketplaces, directories, and ecosystems; representing Phase in Claude, Codex, Cursor, MCP, and other agent environments; GitHub and developer distribution; documentation and onboarding; finding new distribution channels; growth in skill usage and feedback from users. 4. Product development Understand which skills agents actually need, which get used most, and what to build next. Watch usage, quality, user scenarios, and how skills bring new users and activity into Phase. 5. Team adoption Make sure everyone on the team works through AI agents as effectively as possible: tooling, practices, prompts, internal skills, and onboarding. You're the person people come to with: "How do I do this properly with AI?" 6. Internal agent platform Build the internal infrastructure: which agents and skills are in use, where they live, how they deploy, what they cost, how we measure quality, and what access they have. 7. Growing the function As the work grows — build your own AI team and lead it. What matters: You live in agent tools every day. Claude Code, Codex, Cursor, MCP, skills, and your own workflows are how you work. You can build skills and tooling for agents, not just use what already exists. You work agent-first. You can brief an agent, give it the right context, and verify the result. You've shipped something public. An agent, skill, MCP server, plugin, open-source tool, or another project other people use. You think like a product owner, not only an engineer. You understand who the product is for, how to distribute it, and how to measure usage. You can explain and teach. Half the role is helping other people work in a new way. Blockchain knowledge is a plus. Willingness to get up to speed fast is not optional. English good enough for documentation, GitHub, and the international AI community. What done looks like after three months: There's a clear roadmap of Phase skills for AI agents. New skills shipped, or existing ones substantially improved. Phase skills are published and present in the main relevant marketplaces and agent ecosystems. There are clear usage metrics: which skills are used, how often, and for what. The whole team works through agents using shared practices. There's an internal base of Phase agents and skills. There's a short AI strategy for the next six months, agreed with the CTO. How to apply: In your cover letter , please send us a link to something you've already built with agents or for agents — a skill, MCP, agent, plugin, or another project — plus a couple of paragraphs on how you'd approach your first three months with us.