A Week in the Life of a Forward Deployed Engineer (10 clients, 50 hours)

Author here. I’ve been a software engineer for about 12 years, but only more recently in a Forward Deployed–style role embedded with customers. This post is a diary of a typical week from a few weeks ago at an AI startup, working with 10 active clients.

The week included: – Roughly a 50/50 split between meetings and building (coding, prompt work, debugging) – Heavy context switching across Slack/Teams, Linear, Notion, docs, etc. – Unplanned firefighting when an upstream provider (Cloudflare) went down and all our client systems felt it

Happy to answer questions about: – How this work compares to “classic” SWE, solutions engineering, or consulting – What skills actually matter day to day in this kind of role – Tradeoffs: chaos, context switching, and owning both scoping and delivery – When it makes sense for a company to introduce an FDE‑type role at all

I’ve been working as a Forward Deployed Engineer at an AI startup, embedded directly with customers. Last week I juggled 10 clients: parallel builds, Cloudflare outages, pre‑sales calls, and a 50/50 split between meetings and “real” building. This post is a blow‑by‑blow of that week: planning, context switching across Slack/Teams and internal tools, prompt‑tuning an insurance pipeline, firefighting when upstream infra breaks, and feeding patterns back into the product. It’s meant as the honest version of what this role actually looks like, not a recruiting brochure.

thanks for the insights! do you enjoy the job more than the SWE role you had previously?