AI Software Lead

First Student is North America's leading provider of student transportation, helping millions of students get to and from school safely each day. Our technology teams build and support the operational, safety, customer, and employee-facing systems behind that work, including our HALO platform. 

AI is an area of active investment, and this role is part of how we are building practical, reliable, and governed AI-enabled capabilities for the business. 

About the Role 

This is not a traditional software engineering role, and "Lead" does not mean managing people. You will lead agent-driven software delivery: deciding what to build, decomposing work for AI coding agents, supervising their output, validating quality, and owning what ships. 

You will not primarily be measured by writing code line-by-line. You will be measured by your ability to direct AI coding agents effectively, create the scaffolding that keeps their work aligned, review their output with senior-level judgment, and ensure the resulting software is reliable, maintainable, and production-ready. 

When the solution itself is an AI agent - such as a routing assistant, dispatcher copilot, or parent communications agent - you will help design, build, evaluate, and operate it in production. 

You will work primarily on net-new tools and agents, with some work in existing systems. You will build in our standard stack: React, React Native, and AWS. Many tools may touch student data, so FERPA and enterprise governance are part of the work. You will be a peer in a small AI pod and will help shape the AI governance and delivery practices we are building now. 

How You'll Work 

You will operate AI coding agents - such as Claude Code, Codex, Cursor, and successor tools - as a primary means of software production. Your value is not typing speed. It is the judgment to know what to build, how to constrain the agent's work, when the output is wrong, and how to prove the result is right. 

A key part of the role is building the scaffolding around agent-driven development: project context, tests, architectural rules, file conventions, allowed dependencies, review checkpoints, and patterns that reduce drift over time. 

A typical week may include: 

• Translating business problems into specifications, constraints, and implementation plans AI coding agents can execute. 

• Running agents in parallel, reviewing output, and integrating work into maintainable systems. 

• Designing tests, project context, architectural rules, and review checkpoints to keep agent output aligned. 

• Building evals, guardrails, and monitoring for AI-built and AI-embedded systems. Working with stakeholders, architecture, security, and AI governance to clarify requirements and manage risk. 

Key Responsibilities 

Lead agent-driven software delivery 

• Translate business problems into clear specifications, constraints, and implementation plans for AI coding agents. 

• Direct AI coding agents to design, build, test, and ship applications, workflows, and AI-enabled tools. 

• Decompose ambiguous requests into agent-executable work and validate that outputs meet business and technical requirements. 

• Run agents in parallel where useful; review, reconcile, and integrate their output. 

• Prototype rapidly to gather feedback and inform product direction. 

Maintain quality, reliability, and architectural discipline 

• Design and maintain agent harnesses, including project context, architectural rules, file conventions, allowed dependencies, and review checkpoints. 

• Use test-driven practices to constrain agent output and catch drift early. 

• Ensure appropriate automated test coverage across unit, integration, end-to-end, and contract tests based on risk. 

• Review agent-produced code for correctness, security, maintainability, and architectural fit. 

• Enforce patterns that keep codebases maintainable as AI agents contribute to development. 

Build and operate AI-enabled systems 

• Build solutions using design systems, tool use, APIs, function calling, RAG, DAG, MCP, context engineering, harness engineering and multi-agent workflows where appropriate. 

• Evaluate models and AI tools across providers based on cost, quality, latency, reliability, security, and fit for purpose. 

• Implement evals, monitoring, logging, and guardrails so AI systems are measurable and supportable.