About The Position

Fable is hiring a Software Engineer with hands-on AI experience for a 4-month, full-time fixed-term contract, with the possibility of extension based on project needs. The engineer will design and build AI integrations using the Model Context Protocol (MCP) and related technologies across Fable’s platform and APIs. Expected outcomes include production-ready AI integration features, secure customer-specific authentication and access, evaluations and monitoring for AI tool use, and setup guides and technical documentation. This is a hands-on role focused on taking ownership, moving quickly, and delivering production software. The position reports to the Director of Engineering and works closely with Product Management, senior engineers, and Quality Engineering. Fable is remote-first; the team works regular business hours Eastern time, Monday to Friday. The engineer may occasionally attend in-person collaboration at the Toronto office. Preference is given to candidates in the Greater Toronto Area, though candidates elsewhere who can travel for meetings are welcome. Fable works with global enterprises to make products more accessible for people with disabilities. Its customers include Walmart, Slack, and Shopify. The company describes its environment as collaborative and mission-driven, focused on building a more inclusive digital world.

Requirements

  • At least 3 years of professional backend development experience building and running production services and APIs.
  • Strong skills in JavaScript or TypeScript and Node.js.
  • Experience designing, building, and integrating REST APIs in production.
  • Experience with automated testing, Git, code review, and CI/CD, such as GitHub Actions.
  • Clear communication and the ability to work independently and deliver at a fast pace.
  • Practical, customer-focused problem-solving skills.
  • Hands-on experience building AI tool integrations or agent tools.
  • Experience integrating large language models (LLMs) into applications.
  • Experience with tool calling (function calling) and structured outputs.
  • Experience with prompt engineering and writing tool descriptions that models use reliably.
  • Experience testing and evaluating AI behavior.
  • Experience using AI-assisted software development.

Nice To Haves

  • Working knowledge of OAuth 2.0 or 2.1, token-based authentication, and access scopes.
  • Experience with Express.js and the Serverless Framework.
  • Experience with AWS services such as Lambda, ECS, or API Gateway, and with Docker.
  • Experience with MongoDB.
  • Experience with Auth0 or similar identity providers.
  • Experience with AI agent frameworks and SDKs.
  • Experience with JSON-RPC or other real-time communication protocols.
  • Experience with Infrastructure as Code.
  • Experience with observability and monitoring platforms.
  • Experience building integrations with third-party SaaS tools.
  • Knowledge of web accessibility (WCAG) or experience with assistive technology.

Responsibilities

  • Design and build MCP servers and AI integrations using the official TypeScript MCP SDK.
  • Build MCP tools, resources, and prompts that make platform data and actions available to AI assistants and agents.
  • Implement remote Streamable HTTP transport and local stdio transport as needed, in line with the current MCP specification.
  • Connect MCP tools to existing APIs and services, handling data mapping, pagination, retries, idempotency, and rate limits.
  • Keep tool versions and schemas stable as the server evolves.
  • Design a small, organized set of tools around user needs rather than creating one tool for each API endpoint.
  • Write clear tool names, descriptions, input and output schemas, and error messages that AI models can act on.
  • Build evaluations to test how reliably AI models select and call tools, and use results to improve tool design.
  • Test integrations across MCP clients, including with MCP Inspector.
  • Prototype AI use cases with Product and develop the strongest ones into production features.
  • Implement authentication and authorization for MCP servers, including OAuth 2.1, to ensure agents can access only data their users are permitted to reach.
  • Treat tool-returned content as untrusted and guard against prompt injection and overly broad permissions.
  • Log which agent called which tool without exposing sensitive customer data in logs or responses.
  • Write clean, well-tested code, including unit and integration tests, and participate in code reviews.
  • Deploy and run services on AWS through CI/CD pipelines.
  • Monitor tool latency, failure rates, and usage in addition to overall server health.
  • Write setup guides and technical documentation.
  • Build with accessibility in mind so delivered features work well for people using assistive technology.
  • Use AI-assisted development tools while maintaining quality.
  • Own work from design through production in collaboration with Product Management, senior engineers, and Quality Engineering.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service