AI Engineer

Heysurface, Inc.; United States; Remote; USD 100,000 - 180,000; Full time

Benefits offered by this job

Early-stage equity

Job summary

Heysurface, Inc. is looking for the first engineer to build the core of an AI product that tracks brand appearances in answers from major AI assistants and search overviews.

You will shape the technical direction, define what to build, and create integrations with multiple AI providers to collect responses. You will design the scoring and aggregation, draft changes for approval, and develop an evaluation harness to test prompts before deployment, while enabling per-customer memory and spending

Qualifications

• Experience shipping products that call model APIs and handle rate limits, partial failures, and usage costs.
• Strong TypeScript skills and willingness to write maintainable code.
• Use reproducible evaluation sets to inform decisions; avoid unreproducible benchmarks.

Responsibilities

• Build and maintain direct integrations that collect responses from five AI providers.
• Develop scoring and aggregation that turn collected answers into a clear position.
• Create the workflow for proposing and drafting changes, with human approval before execution.
• Build an evaluation harness to assess prompt changes before they ship.
• Move work into background processing, and develop per-customer memory and spending limits.

Skills

TypeScript

APIs

Product mindset

Experimentation

System design

Job description

Role overview

As the first engineer, you will build the core of an AI product that tracks whether brands appear in answers from major AI assistants and search overviews. The system identifies opportunities to improve those answers, drafts proposed work for approval, carries out approved changes, and measures results. The role includes shaping the technical direction and defining what to build.

Responsibilities

• Build and maintain direct integrations that collect responses from five AI providers.
• Develop scoring and aggregation that turn collected answers into a clear position.
• Create the workflow for proposing and drafting changes, with human approval before execution.
• Build an evaluation harness to assess prompt changes before they ship.
• Move work into background processing, and develop per-customer memory and spending limits.

Requirements

• Experience shipping products that call model APIs, including handling rate limits, partial failures, and usage costs.
• Strong TypeScript skills and a willingness to write maintainable code.
• Use reproducible evaluation sets to make decisions; treat unreproducible benchmarks cautiously.
• Sound judgment about when an AI agent should ask for input and when it should act.
• Comfort defining the product and technical plan without a fully written specification.

Benefits and work setup

• Early-stage equity is offered; its value is uncertain.
• The product has not launched and has no revenue yet. The stated public launch target is September 2026.