Staff AI Engineer

TaskRay•,

1d•$163,000 - $246,000•Remote

About The Position

TaskRay is seeking a Staff AI Engineer to lead the technical design of its AI platform. This role is central to the company's next 18 months, focusing on an off-platform agent layer that integrates deeply with their Salesforce-native distribution. The position involves owning the architecture, setting engineering standards for LLM development, and partnering with senior leadership to ship the product. The role also includes customer visibility through design partner engagements, where the engineer will be the technical voice in virtual meetings. The AI agents will be developed sequentially, starting with internal-facing capabilities (PM Agent and Execution Agent) before expanding to an External Onboarding Agent. The MCP server will serve as the foundational layer for all these agents. This is a hands-on builder role where technical judgment will significantly influence the work of other engineers on the team.

Requirements

Track record of shipping production agent systems at meaningful scale (real users, real reliability requirements, real failure modes you have debugged).
Deep, hands-on experience designing agent systems in production: tool use, multi-step reasoning, retrieval pipelines, and the evals required to iterate on them safely.
Strong Python proficiency (the lingua franca of the AI stack), with production experience deploying real systems.
Hands-on experience with Anthropic Claude or comparable LLM provider APIs at production scale, including reliability patterns, eval harnesses, and cost and latency engineering.
Demonstrated technical leadership: scoping work for multiple engineers, making and defending architectural decisions, and partnering credibly with senior stakeholders.
Comfort in customer-facing technical discovery and design partner contexts. You can listen well, scope tightly, and hold your own with a senior buyer.

Nice To Haves

Experience with Model Context Protocol (MCP), agent orchestration frameworks, or comparable tooling.
Experience operating in compound or agentic development workflows (Claude Code, Cursor, or equivalent), with a point of view on how teams should adopt them.
Background that includes one or more of: customer-embedded technical engagements, ML platform, or research-to-production transitions.

Responsibilities

Own the architecture of TaskRay's off-platform AI tier, including the contracts that connect it to our Salesforce-native core.
Make and defend build versus buy decisions across the agent stack: orchestration, retrieval, evaluation, observability, and guardrails.
Partner with our Architect and senior Salesforce engineering lead on the data contract between the Salesforce platform and the off-platform AI layer.
Translate the company's AI strategy into a buildable technical plan.
Architect and lead the build of our MCP server, the platform layer that connects agents to TaskRay's project, customer, and onboarding data and gives us control over usage analytics, packaging, and external agent connections.
Design and lead the build of our PM Agent, an internal-facing agent that materially increases customer team productivity by maintaining live project status and surfacing the next best action.
Design and lead the build of our Execution Agent, an internal-facing agent that completes tasks on behalf of human counterparts inside TaskRay workflows.
Design and lead the build of our External Onboarding Agent, the customer-facing agent that handles status updates, document collection, and end-customer interaction across the onboarding lifecycle.
Build the eval, observability, and feedback infrastructure that lets us iterate on agents safely against real customer data.
Set our approach to retrieval, tool use, and multi-step reasoning, aiming for production-quality systems.
Partner with design partner customers to discover repeatable workflow patterns worth productizing.
Lead the technical scoping for design partner work.
Be the technical voice in customer engagements alongside Product.
Establish TaskRay's engineering standards for AI feature development: prompt iteration, eval-driven workflows, agent observability, cost and latency budgets, security and compliance posture.
Actively contribute to our agentic coding standards and norms.
Mentor senior and mid-level engineers contributing to AI work, and shape the broader hiring bar for AI capability on the team.