The Role

We are hiring a Senior AI Engineer to build production agent capabilities for TaskRay's AI platform. You will work closely with our Staff AI Engineer and our product team to ship features that customers actually use, against the most demanding bar in our company: agents that automate real implementation work for enterprise onboarding teams.

You will work across the full AI stack: prompt and agent design, retrieval, evals, tool integrations, and the off-platform service layer that hosts it all. You will partner with design partner customers to learn what the agent needs to do before generalizing it into the product.

We are sequencing these agents deliberately, starting with internal-facing capabilities (PM Agent and Execution Agent) before extending to the External Onboarding Agent. The MCP server is the foundation that underpins all three.

This is a heads-down builder role with leverage. We are small, the work is real, and shipping speed matters.

Note

We are hiring AI engineers, not Salesforce engineers. You bring the agent craft (production LLM systems, evals, retrieval, tool use, multi-step reasoning). We will teach you what you need to know about the Salesforce platform our product is built on. The integration work is real but it is not the center of the role; the agents are.

What You'll Do
AI Feature Development

• Build production agent capabilities for our PM Agent (the internal-facing agent that maintains live project status and increases customer team productivity), our Execution Agent (the internal-facing agent that completes tasks on behalf of human counterparts), and our External Onboarding Agent (the customer-facing agent that handles status, document collection, and end-customer interaction).
• Contribute to our MCP server, the platform layer that connects these agents to TaskRay's project, customer, and onboarding data
• Design prompts, agent loops, and tool integrations that meet customer-grade reliability, not demo-grade
• Own subsystems end to end: retrieval pipelines, eval harnesses, prompt libraries, tool registries, and the observability around them

Design Partner Work

• Partner with design partner customers to ship custom agent workflows that teach us what to productize. Engagements run via video and shared tooling, not on-site
• Translate customer-specific work into reusable building blocks for the core platform
• Partner with Product on customer-facing demos, walkthroughs, and feedback loops

Platform and Integration

• Help build and maintain the MCP server and the off-platform service layer that hosts our agent capabilities
• Develop integrations with adjacent systems where the agents need them (Google Drive, calendar and meeting transcripts, document stores, CRM data)
• Contribute to the eval, observability, and reliability infrastructure our Staff engineer is establishing

Engineering Craft

• Write clean, well-documented code that your teammates will thank you for
• Participate in code reviews with a constructive, growth-oriented mindset
• Actively contribute to our agentic coding standards and norms

What We're Looking For

• You have shipped LLM-based features to production users. You have lived through the gap between a working demo and a system that actually serves customers
• Hands-on experience with retrieval pipelines, agent loops with tool use, and the evals that keep them honest
• Strong Python proficiency (the lingua franca of the AI stack), with production experience deploying real systems
• Hands-on experience integrating LLM provider APIs (Anthropic Claude, OpenAI, or equivalent) into real applications, including evals, prompt iteration, and cost and latency engineering
• Strong written communication. You can explain technical decisions clearly to product, customer success, and customers themselves. We are level-agnostic on years; we care about what you have shipped.

Strongly Preferred

• Experience with Model Context Protocol (MCP) or agent orchestration tooling
• Curiosity about the Salesforce platform. You do not need to know it. We will teach you. But you should be excited about integrating agents with the system of record where most enterprise customer-facing work actually happens.
• Experience with compound or agentic coding workflows (Claude Code, Cursor, or equivalent)
• Comfort in customer-facing technical contexts, including design partner work and customer demos (all virtual)

Compensation

Cash compensation for this role is commensurate with experience. The estimated salary range is between $148,000 - $209,000 per year. This position is also eligible for bonus based on company and individual performance targets.

Additionally, TaskRay offers highly competitive non-cash compensation including:

• Medical, dental, and vision benefits
• Every other Friday off and a team that respects your time outside of work
• Flexible PTO
• 12 weeks paid family and medical leave, 16 weeks for birthing people
• Vacation bonuses
• Anniversary bonuses
• Company-paid life insurance
• 401(k) matching
• Cell phone reimbursement stipend
• Employee Assistance Program