Selected work · 02

Architecting the FalconX custody platform

A greenfield custody platform creates a rare opportunity and a serious responsibility. I worked from hands-on prototypes toward reusable foundations that could make each new asset and product safer to support.

Institutional custody · Transaction infrastructure

Visit FalconX

My role

Hands-on architect and builder of a greenfield custody platform, from early prototypes through reusable foundations.

Context

The conditions that shaped the work.

Context
A new institutional custody platform needed foundations for the next wave of assets and products rather than an extension of a single existing integration.
Constraints
Transaction construction, signing, policy, indexing, broadcasting, balances, and staking carry unforgiving failure modes when the system protects customer assets.
Core tension
The platform needed enough common structure to create leverage without hiding the chain specific behavior that determines correctness.

Timeline

How the work progressed.

  1. 01

    Prototype

    Make the hard path real

    Hands-on prototypes exposed the transaction and signing concerns that a credible platform would need to own.

  2. 02

    Foundation

    Identify the repeating responsibilities

    The work separated durable platform concerns from the asset specific behavior that could not be abstracted safely.

  3. 03

    Capability

    Make the next asset safer

    Reusable, event-driven foundations turned support for new assets into a repeatable platform capability.

Decisions

Key choices and their tradeoffs.

Prototype before fixing the platform shape

Prove the direction through working transaction paths before treating the first architecture as permanent.

Tradeoff

Some early work optimized for learning, but it reduced the risk of generalizing around imagined requirements.

Abstract recurring responsibilities

Create reusable foundations for transaction construction and the surrounding event flow.

Tradeoff

The platform had to preserve asset specific behavior rather than force every chain into a false common model.

Measure leverage through the next launch

Judge the abstraction by whether the next asset became safer and more repeatable to support.

Tradeoff

Generality alone was not the goal. Each shared boundary needed to remove real launch risk.

Decision path
  1. Product capability
  2. Transaction construction
  3. Policy and signing
  4. Broadcast and state

Outcomes

What changed.

  1. Established reusable, event-driven foundations for the custody platform.
  2. Turned support for new assets into a repeatable platform capability.
  3. Created a foundation that product teams could build on without flattening asset specific risk.

Lesson learned

The useful abstraction is not the most general one. It is the one that makes the next launch safer.

Continue

Next: BitGo

Building the organization that growth demanded.