Senior Software Engineer - AI Builder

Overview

Pathfinder is Intuit’s intent-based AI test automation platform — natural-language prompts executed by autonomous agents against live applications, with 1M+ E2E runs recorded and 18+ teams building on it across TurboTax and beyond. We’re moving from a proven platform into the next phase: PRD-to-test generation, Figma-to-test generation, and full SDLC integration for auto-generated tests. This is greenfield problem space, not a known pattern with reference implementations to lean on.

We’re looking for a Senior Software Engineer to own the layer that determines whether Pathfinder scales: multi-agent execution orchestration and cross-platform execution caching. This is the kind of person who treats an LLM rate limit not as a blocker to escalate but as a system design problem to solve — someone fluent in building with AI, not just building AI features, who can reason about agentic execution patterns, cost/performance tradeoffs, and orchestration architecture with no existing playbook to reference. You’ll be hands-on 80-90% of the time, architecting and shipping the infrastructure that lets autonomous agents run reliably at scale, while setting technical direction for a small team solving problems nobody has solved before.

This role suits a builder — someone who prototypes fast, is comfortable being wrong in week one and right by week three, and treats ambiguity as the interesting part of the job rather than a reason to wait for more clarity.

Responsibilities

• Design and build the multi-agent execution orchestration and cross-platform execution-caching infrastructure that Pathfinder’s scale and LLM cost efficiency depend on
• Drive significant technical initiatives end-to-end, from functional requirements through technical specification, test planning, and production rollout
• Act as the technical subject matter expert on agentic execution systems — lead a small team through hard, unprecedented programming and design problems
• Prototype and validate proofs of concept quickly; use AI-assisted development tools fluently as part of your own workflow, not just as a subject you build for
• Generate technical documentation that communicates architectural and design tradeoffs clearly to both engineering teams and business stakeholders
• Resolve defects across QA, pre-production, production, and post-release patches for a system running 1M+ E2E test executions

• Work cross-functionally with product management, QA/QE, and partner teams (18+ teams currently building on Pathfinder) to unblock their adoption and surface platform gaps early

• Operate in Agile/Scrum cadences while retaining the independent judgment to make calls the backlog hasn’t anticipated yet

Qualifications

• Bachelor’s degree in Computer Science, Engineering, or equivalent practical experience; 3+ years building cloud-based platforms or services
• 3+ years developing systems/software for large-scale environments (Java, Springboot), with demonstrated ability to design for scale quickly rather than needing a decade to build the instinct
• Hands-on experience building with AI: you’ve shipped something using LLM APIs, agentic frameworks, or AI-assisted coding tools as a core part of the system, not as an add-on. Comfortable reasoning about prompt-driven execution, model behavior under load, and the failure modes that come with non-deterministic systems
• Fluent daily user of AI-assisted development tools (Claude Code, Cursor, Copilot, or similar) — this role expects you to build faster because of AI, and to bring that instinct to designing infrastructure where AI agents are the primary actor
• Practical experience with microservices and messaging technologies (REST, GraphQL, Kubernetes, Kafka, ActiveMQ) and database design across SQL (MySQL) and NoSQL (DynamoDB, Cassandra)
• Some exposure to the cost/performance mechanics of running LLMs in production — rate limits, token economics, caching strategies, orchestration patterns — is a strong plus even without years of tenure in the space, since almost nobody has years of tenure here yet
• Strong CS fundamentals: data structures, algorithms, and system design intuition that holds up with no reference architecture to copy
• Working knowledge of TDD and SDLC practices, with the judgment to know when rigor matters and when speed does
• A builder’s bias toward shipping working prototypes over long design docs, balanced with the discipline to document architecture decisions clearly once…