This position is no longer accepting applications.
It stays published for reference. Browse open web3 jobs or similar roles below.
AI Developer Tools Engineer
Input Output website
Remote (UK)
1 month ago
full-time
closed
Core Responsibilities
Enabling Agentic Development on Midnight
- Own and improve the tooling, context files, SDKs, and documentation that let AI coding agents (e.g. Claude Code, Cursor, Copilot-style agents) build on Midnight effectively
- Design, build, and optimize agent harnesses and orchestration workflows for Midnight-specific development tasks
- Identify and close gaps where agentic tooling fails, hallucinates, or produces incorrect Midnight-specific code
- Build and maintain reference implementations, prompts, and examples that improve agent output quality
- Evaluate new agentic coding tools and protocols (e.g. MCP servers) for relevance to the Midnight developer ecosystem
Builder & Agentic Tooling Feedback Loop
- Monitor how builders and agentic tools are actually using Midnight — successes, failure modes, and workarounds
- Synthesize feedback from agentic tooling usage and community builders into clear, structured input for the core protocol roadmap
- Maintain a regular reporting cadence to Product and Engineering on developer/agent experience trends and blockers
- Represent the AI-assisted builder perspective in roadmap and prioritization discussions
AI Dev Tooling Maintenance
- Own ongoing maintenance, accuracy, and reliability of Midnight Expert and other AI-enabled developer tools for the community
- Keep tool knowledge bases, embeddings, and reference content current as the protocol evolves
- Triage and resolve community-reported issues with AI dev tooling
- Define and track quality metrics (accuracy, usefulness, adoption) for AI-enabled dev tools
Recommended Experience & Qualifications
- Hands-on experience building, integrating, or maintaining AI coding assistants or agentic dev tooling (e.g. Claude Code, Cursor, GitHub Copilot, MCP servers, RAG-based coding assistants)
- Practical understanding of how AI coding agents consume context — prompting, retrieval, tool use — and what makes them succeed or fail on a given codebase or protocol
- Familiarity with smart contract or protocol-level development (Cardano/partner-chain, ZK-based systems, or equivalent) is a strong plus, though not mandatory at this level
- Experience turning qualitative user or usage feedback into structured product or roadmap input
- Strong technical writing skills — this role will shape documentation and reference material AI agents and humans both rely on
- Comfort working async and independently across a wide working-hours band (UTC-5 to UTC+5)
- Bonus: built an application on Midnight or another privacy network