AI Agent Developer

We are seeking a skilled Backend Engineer on a contract basis to develop a persistent, stateful AI agent that manages intake workflows across Slack and Notion. This is not a typical chatbot but an always-on process designed to reason, wait, follow up, and close the loop much like a human manager. The role involves building a robust system that integrates with Slack and Notion while leveraging the Claude API to create an intelligent, autonomous workflow manager.

Key Responsibilities

- Develop a long-running backend process using Node.js, TypeScript, and Python, incorporating Slack Bolt with Socket Mode for event-driven and scheduled operations.
- Implement idempotent and restart-safe state management to ensure no loss of context or duplicate actions occur.
- Build a Claude API tool-calling loop where the AI model determines what actions to take, and a decision engine controls when those actions execute.
- Utilize Notion as both the system of record and durable memory for the AI agent, ensuring reliable data persistence.
- Design configuration-driven behavior, including escalation ladders, business hours, and voice settings, enabling easy redeployment for new clients without code changes.

Required Qualifications

- Minimum of 3 years of experience developing production backend systems using TypeScript and Node.js.
- Hands-on expertise with Slack Bolt, including Socket Mode, threaded replies, direct messages, and Block Kit components—not limited to basic webhook usage.
- Practical experience working with the Notion API, including databases, formula and relation properties, and the newer multi-data-source model.
- Demonstrated knowledge of large language model (LLM) tool-calling or agentic loops, particularly with Claude or OpenAI function calling, understanding the critical distinction between model-driven decisions and code-driven timing.
- Strong background in designing for idempotency and crash recovery, employing techniques such as optimistic locking and write-before-side-effect patterns. This is essential for the role.
- Proven ability to work asynchronously and independently in a remote environment, managing priorities and meeting deadlines effectively.
- Bachelor’s degree in Computer Science, Engineering, Data Science, or a related field, or equivalent practical experience.

Preferred Qualifications and Benefits

- Experience with Lawmatics, legal intake workflows, or SLA/exception-based automation is a plus.
- Prior work building AI agents or automation solutions in SaaS, logistics, transportation, or other physical operations domains will be advantageous.
- Strong familiarity with AI/ML tools and frameworks, including prompt engineering and conversational AI platforms, with the ability to design and optimize agent behavior.
- Solid understanding of data structures, databases, and system design principles to enable secure and efficient retrieval, processing, and action on operational data.
- Experience in product or workflow design, including gathering requirements, mapping user journeys, and translating business processes into automated agent flows.
- Excellent analytical and problem-solving skills, with a focus on experimentation, debugging complex behaviors, and iterating based on feedback.
- Effective communication skills, both written and verbal, to document solutions clearly and collaborate with technical and non-technical stakeholders.

This is a contract engagement with a fixed price and milestone-based payment structure, progressing from a proof-of-concept loop (v0) to production-ready deployment (v1). The position is fully remote and supports asynchronous work styles.

If this opportunity aligns with your expertise, please be prepared to share portfolio examples of Slack or Notion integrations you have delivered, especially those involving state persistence across restarts.