Senior Agentic Engineer - Product​/UX

All of our code is written through our AI process. We adhere to the Rails Way on a modern Rails 8 stack with Hotwire and Hotwire Native.

Our agentic development pipeline ships features quickly, with review informed by our codebase, business logic, and personas. Our custom AI process produces product specs, code, documentation, and tests. This isn't "code with AI's help". It is a structured, end-to-end engineering workflow where AI agents plan, execute, and validate under human oversight.

The best fit for our engineering team has enough coding experience to bring great judgment. Someone who recognizes well-structured code, thinks architecturally, and takes a "trust yet verify" approach to AI output, especially because our customers and industry have strong needs around sensitive information.

You'll work with engineering leadership and our AI team to develop a holistic understanding of the platform, then own the full development cycle for the work you ship. You'll also apply your engineering mindset to the agentic development process itself to help it evolve.

AI Tooling: Claude Code
· Wispr Flow

Context: AI-native engineering workflow
· HIPAA-compliant legal tech

The focus of this role is improving the user experience of our application, improving how our agentic development process delivers product quality, and improving AI's ability to assess and enhance usability.

We need someone who lives at the intersection of product thinking and technical validation — someone who can take an idea from concept to shipped feature, and who knows the difference between "it works" and "it works well for the people using it." This takes the skills of a team lead who can plan and delegate to agents holistically and strategically.

You'll orchestrate how they build features, validate whether the output actually solves user problems, and continuously improve the product (and process) through their work.

You need to understand the technical implications of a modal vs. full-page, synchronous vs. asynchronous, monolith vs. component. You need to be able to look at a generated feature and say "this flow is confusing" or "this doesn't actually solve what attorneys need" — and also look at the code behind it and know if it's well-structured.

Over time, you'll find new ways to use AI to improve product quality itself — dogfooding the app, surfacing friction points, measuring how many clicks a workflow takes, and suggesting improvements. You'll improve the AI's own understanding of UX/UI best practices so it produces better output by default. You'll build the product observability and feedback loops that let us iterate faster and smarter.

• Use and continuously improve our existing agentic orchestration to build and iterate on features end to end. Plan the work, direct the agents, review the output, and improve the process
• Learn our codebase and domain, and help tighten and refine the platform as it exists today
• Take new product ideas and feature requests from concept to implementation plan to deployable code — creating specs through voice-first, AI-assisted workflows, not traditional writing
• Implement and manage product observability: track user behavior, click paths, time-on-task, drop-off points
• Use AI tools to dogfood the app — surface areas where things look difficult, take too long, or require too many steps, and suggest improvements
• Improve AI's ability to test the platform for usability and troubleshoot technical issues
• Work with our personas and user research to ensure features serve the people who actually use the product — attorneys, CASA volunteers, caregivers, families

• Product adoption and stickiness — partner with stakeholders to drive rollout, measure adoption, and iterate until features become default workflows
• End-to-end user flow testing — automation across web, iOS, and Android that validates the experience, not just the code
• Product observability — dashboards and tooling that show how users actually interact with what we've built
• Feature polish and iteration — taking shipped features from "works" to "delights"
• Product documentation (memory) — AI-generated and…