The Role
Engineers on the AI Developer Foundations will build the tools that build the product. It's a new team that owns how AI gets used inside engineering, product, and design at SimplePractice: the conventions for using these tools, the context agents read before they touch our code, the bar generated work has to clear before a human spends time on it, and the training that makes AI a useful tool to all of engineering.
This is a hands-on role reporting to the VP of Technology. You'll write code most days, mostly in and around a large Rails monolith, and you'll also write the conventions other engineers follow.
Responsibilities
- Own the context layer agents read: rules files, docs conventions, grounding data, with CI and named owners keeping it current
- Publish the standards for how we prompt coding agents, how we supply context, and what a spec needs to contain before it's worth handing to an agent
- Define the review gates any AI-generated artifact clears before a human spends time on it, and get what reviewers catch flowing back into the conventions
- Take the prototypes running today and turn the ones that hold up into supported workflows that run for someone other than their author
- Evaluate harnesses, models and vendors against criteria and eval sets
- Work the product-to-engineering handoff with PMs, designers and EMs
- Teach. Documentation, worked examples, office hours, and time sitting with a squad while they work with the tools we build
- Help us draw the boundaries for where an agent runs unattended, where a human signs off, and where we don't use one at all
Desired Skills & Experience
Education & Experience
- BS/MS in Engineering, Computer Science, or related field, or equivalent experience
- 7+ years building production software and then maintaining it. You’ve shipped something, lived with the decisions, and found out whether a design decision held up
- Strong problem-solving and communication skills; comfortable in fast-paced, cross-functional environments
Ruby on Rails
- Recent, direct Rails experience in a large codebase.
- Ability to read unfamiliar application code quickly and judge whether what an agent produced in it is any good
AI-Assisted Engineering
- LLM-backed systems you've shipped to users and kept running, past the point where better prompting stops helping
- Experience building context or memory handling, and writing evals
- Judgment about what belongs in code and what belongs in the model
- Track record evaluating tools with evidence
Developer Tooling & Enablement
- Internal tools, platforms, conventions or workflows that other engineers chose to use.
- CLI and harness-shaped software, and comfort connecting systems over APIs