- Job title
- Founding AI Engineer
- Job location
- San Francisco, CA, US
- Job description
- ## About Us
Coach is building the AI coach for the physical world. Starting with in-person sales.
For centuries, field organizations have relied on layers of management to relay information, coach teams, and understand performance. We believe AI will radically change how these companies operate and how people learn and improve at work.
Today, understanding why a sales rep succeeds or struggles still requires being alongside them in the field. That doesn’t scale. And yet roughly 80% of global sales still happen offline.
Coach helps employees learn faster, improve, and close more. Managers get a clear view of what’s happening, where people need help, and why, so they can focus on solving problems.
Our custom AI harness turns field conversations and business context into **sales intelligence**, grounding coaching in what actually happens on the ground. Over time, that accumulated knowledge becomes a **company brain**: a shared understanding of how the business works.
We’re bringing AI to people who have never had a reason to use it, by making it part of the human interactions at the heart of their work. In-person sales is only the beginning
## What you’ll own
* **Our AI harness.** Build how agents gather context, access tools, retain knowledge, deliver and track coaching grounded in real conversations.
* **Evaluation and experimentation.** Create datasets, benchmarks, and feedback loops to measure agent and model quality. Investigate failures, challenge apparent patterns, and test whether improvements translate into customer value.
* **Machine learning models.** Design, train, evaluate, and deploy models for behavioral analysis, performance forecasting, and personalized coaching from dataset construction and feature engineering to validation and monitoring.
* **Production infrastructure.** Build and operate the backend, data pipelines, and services behind Coach. Improve reliability, security, observability, latency, and cost.
* **Our software factory.** Work with the team to build the tools, environments, and verification workflows that let our team and coding agents collaborate effectively.
* **Forward deploying.** You will meet our customers and spend time with them to better understand how to satisfy their needs
## Who you’ll work with
We're a small team in SF and Paris, growing fast. Our CTO has been building software and AI across many industries for the last 10 years. Our CEO and COO are repeat founders who built and ran in-person sales teams selling to pharmacies. We've lived the problem firsthand and built Coach to solve it.
With just two employees (Customer Success + Engineering), we took Coach to **$1.5M ARR in 8 months** — and we're scaling aggressively from here.
## What you bring
* Strong backend engineering skills and experience shipping and operating production systems.
* Hands-on experience building AI applications: context management, retrieval, tool use, orchestration, and evaluation.
* Experience developing ML models on real data and taking them beyond experimentation into a usable product.
* Solid foundations in probability, statistics, and machine learning. You can recognize confounding, prevent data leakage, quantify uncertainty, and choose an appropriate baseline.
* Fluency in Python and SQL, and confidence working with cloud infrastructure.
* An AI-native approach to development, with the judgment to understand and verify what your tools produce.
* Curiosity about the people using your work, and the ability to turn an ambiguous business problem into a focused implementation.
We're not hiring for a title or a number of years. We're hiring for what you can build. Whether you've done this for 2 years or 10, if you've shipped real AI systems into production and can turn an ambiguous business problem into a working product, we want to talk.
## Our current stack
* **Backend:** Python, FastAPI, Pydantic.
* **Agent systems:** a custom harness built with PydanticAI and frontier model APIs.
* **Data:** PostgreSQL and Supabase.
* **Infrastructure:** Google Cloud Run, Cloud Tasks, Docker, and GitHub Actions.
* **AI evaluation and observability:** Braintrust.
* **Voice:** LiveKit, Deepgram, and ElevenLabs.
You’ll help evolve this stack and select the modeling tools that fit the problems we tackle. We value strong foundations and evidence of what you’ve built; experience with every tool listed isn’t required.
SF-based, relocation supported (visa sponsorship available)
