Forward Deployed Engineer (FDE) best practices insights

Vinoo Ganesh, CEO of Kepler, shared best practices for Forward Deployed Engineers (FDEs) on the Latent Space podcast, emphasizing that FDEs must return insights to the platform, not just solve client problems. He cited a 2013 Palantir incident where a system failed due to lack of real-world data, illustrating the importance of on-site presence. The role is often misunderstood, with varying definitions across companies.
Key insights
FDE role is often mischaracterized
Ganesh noted that FDEs at different companies, including Snowflake and Anthropic, have vastly different roles, from pre-sales engineers to metric-driven salespeople. He stressed that the true value lies in being an extension of the product team, using on-site problem-solving to gather insights for platform improvements, not just delivering consulting or outsourcing services.
Real-world data exposure is critical for FDE success
The 2013 Palantir Phoenix project failed because the system was built on second-hand requirements without testing against real production data, leading to a server crash requiring 14TB of memory. Ganesh emphasized that such knowledge cannot be learned from process charts, highlighting the need for FDEs to be present in the field.