AI Tooling Developer (1897) 90-100hr

Location: Toronto, Ontario (Hybrid - three days onsite per week)

Contract Term: 14 months

Hours: 40 hours per week

Hourly Rate: $90–$100/hr

Skills: 5+ years Python, Git workflows, API integration, LLM, Model Context Protocol (MCP), Claude, Enterprise ChatGPT, OAuth, SSO, Agile

Industry: Banking

DESCRIPTION:

The AI Tooling Developer is a hands-on technical role within the AI team, responsible for supporting enterprise assistant platforms and the systems that connect them to organizational data and workflows. The successful candidate will design and maintain custom connectors to internal systems, develop a governed library of reusable skills and prompt assets, establish secure frameworks for exposing internal data, and provide practical technical support to business users across the organization.

RESPONSIBILITIES:

• Design, build, and maintain custom Model Context Protocol (MCP) connectors that provide approved AI assistants with secure access to internal systems.
• Determine the appropriate deployment model - organization-scoped remote connectors or locally run desktop extensions - based on data sensitivity, security requirements, and business needs.
• Implement secure authentication and authorization using OAuth, per-user entitlements, narrowly scoped tools, and read-only access by default.
• Instrument connectors with logging and monitoring, and maintain clear documentation, support procedures, and iterative enhancements as source systems evolve.
• Develop, test, and maintain a vetted library of reusable AI skills that encode organization-specific procedures, templates, and domain knowledge, including investment memo formats, model conventions, and policy checklists.
• Manage the skills repository and its lifecycle, including version control, quality standards, review and approval processes, assigned ownership, deprecation, and controlled distribution to appropriate user groups.
• Build evaluation and regression-testing capabilities to ensure skills and connectors continue to perform reliably as underlying models and platforms change.
• Establish reference architecture, security baselines, and approval pathways for exposing internal data through MCP, in partnership with Information Security, Data Governance, Privacy, and Legal.
• Serve as the organization’s hands-on subject-matter expert for Claude - including Desktop, Projects, Skills, and connectors - and Enterprise ChatGPT, including connectors and agent features.
• Monitor platform releases and translate new capabilities into practical recommendations on what the organization should adopt, pilot, or avoid.
• Lead office hours and working sessions for users across investment and corporate divisions, translating workflow challenges into practical, buildable solutions.
• Partner with the AI adoption and training function by providing technical content, demonstrations, and guardrails for curricula and the superuser network.
• Perform other duties as required.

REQUIREMENTS:

• 5+ years of hands-on Python experience, proficiency in Git-based workflows and API integrations.
• Direct experience developing solutions on LLM large language model platforms, including tool use, agentic patterns, retrieval, and evaluation.
• Experience with MCP is a strong differentiator; comparable experience in enterprise integrations, APIs, or AI tooling may be considered.
• Working knowledge of enterprise security and identity concepts, including OAuth, SSO, access scoping, entitlements, data classification, and audit logging.
• Strong technical writing skills, with the ability to create clear instructions that can be followed reliably by both people and AI models.
• Demonstrated ability to work directly with investment and corporate professionals and translate workflow challenges into practical technical solutions.
• Experience in financial services or another regulated industry is considered an asset.

Note: AI-enabled systems may be used to support the initial screening of applications.

To Apply: https://directitrecruiting.com/job/ai-tooling-dev-1897/