About the role

Build realtime voice agents for mechanics.
Architect AI-powered retrieval systems.
Extend bidirectional sync engine for ERPs.
Create nimble UIs for service work.
Develop evals and monitoring tools for AI.

Skills the posting asks for

Ruby
React
TypeScript
Python
AI

Summarised from the employer’s posting, which is reproduced in full below.

The employer’s full posting

Keycard is building the intelligence layer for the skilled mechanics that keep America running.

Heavy machinery builds the physical world around us. It breaks ground for data centers, paves our roads, and turns seeds into harvest. Every one of those machines depends on mechanics.

Unfortunately, there’s a massive shortage of mechanics, and the gap continues to grow .

We believe voice as a computing interface can transform how mechanics work. We’re building intelligence that meets them where they actually are: driving between jobs, at the machines with grease on their hands, or on a job site miles from the shop.

Keycard works with leading machinery dealers (across Caterpillar, John Deere, Komatsu, Bobcat, and more) to get their skilled mechanics back to doing what they love. Our large customers close jobs 3.2x faster with Keycard.

Our mission is to build the true System of Action that keeps America moving. We’d love your help.

Who We Are

You’ll be working directly with the founders, Jerry Zhang and Arjun Bali .

We previously founded Sodatone , which was acquired by Warner Music . We trained deep learning models to predict future superstars, on terabytes of data we ingested from across the web, and helped discover multiple #1 hits.

We spent 18 months searching for the right market to build in next: one deeply important to the future of America, and deeply overlooked. 50+ conferences and customer visits showed us just how much of the economy depends on these workers, and how little technology has been built around them.

What You'll Do

As a founding engineer, you’ll make core product and architecture decisions, ship across the stack, and work directly with users.

Build realtime voice agents that are equally intelligent and delightful. Agents that deeply understand the world of heavy machinery, while still being able to talk shop.
Architect AI-powered retrieval systems across millions of repairs, bringing context and intelligence to every part of the repair job.
Extend our bidirectional sync engine that moves context between Keycard and customer ERPs.
Build nimble UIs that make complicated service work feel simple, whether you’re in a truck, on the shop floor, or behind a service counter.
Develop the evals, monitoring, and review tools that tell us exactly what our AI captured, what it missed, and whether a change made the product better.
Take on leadership roles as we grow, as a core product and technical contributor.

Tech Stack

TypeScript, Next.js, Ruby, Python, React Native, Expo, Postgres, Redis, AWS, Snowflake, OpenAI, Anthropic, Gemini

What We're Looking For

Experience building LLM products: prompt engineering, evals, tools, retrieval techniques.
Experience with React, TypeScript, Ruby, or Python.
Proven ability to prototype, ask for feedback, and iterate quickly.
Experience owning end-to-end production systems.
Strong product taste. You do not need to be a designer, but you notice the details that make software intuitive, polished, and delightful to use.

Bonus points for:

Prior experience as a founder or founding engineer - you know what it means to balance craft, ownership, and speed.
Previous roles where you worked directly with customers or led technical projects with external stakeholders.

Logistics

We’re based in New York City. We prefer to work together in person but are open to remote for the right fit.
Generous budget for the AI tools that help you do your best work.
Meaningful early-stage equity.