How Palantir Scaled: Why the Best Software Is Built Backwards

Video summary

Palantir architect Akshay Krishnaswamy explains how field engineers turn customer pain into products and scale software backwards.

On the a16z podcast, Palantir Chief Architect Akshay Krishnaswamy describes forward-deployed engineering (FDE) as building backwards from a real-world outcome: engineers solve a customer’s specific problem in the field, then generalize what works into the product. Unlike services roles tied to hours or contract deliverables, FDEs are accountable for outcomes and for improving the platform—the job, as Palantir’s CTO jokes, is to “absorb pain and then excrete product.” Krishnaswamy says effective FDEs combine technical range, resilience, and the ability to turn difficult customer encounters into product insight. Early teams should stay tightly focused, treating field engineering as a way of building the product, not a separate function. As Palantir scales, its platform should increase engineers’ leverage rather than reduce ambition. Custom work may lead to new platform features, tailored integrations, or trained partners; there is no universal rule. The model requires the whole organization—engineering, product, and customer-partnership teams—to coordinate and support field teams.

Chapters

1. 0:00Introduction: Akshay Krishnaswamy’s 2012 move from biomedical engineering to Palantir
2. 2:17Defining FDE: Build backward from field outcomes, with 6–12 months of field rotation
3. 4:49FDE vs. Other Roles: Outcome-based work turns customer pain into product
4. 6:09Building and Managing Teams: FDE traits, graduate-school dropouts, and tough interviews
5. 9:55Challenges and Evolution of FDE: Foundry, Gotham, and scaling high-autonomy teams
6. 15:27Maintaining Product Focus: Ambitious goals, custom integrations, and organization-wide support