01
Understand
Reverse-engineer the real constraints and failure modes.
Staff systems engineer · from unknown system to operating model
The hard part is rarely the code.
I find the boundary that makes the whole system make sense.
I work where operating systems, browsers, devices, cloud services, and AI meet. I trace failures across boundaries, create the abstraction the system was missing, and turn one-off technical wins into platforms teams can operate and extend.
How the work compounds
02
Operate
Build observability, automation, and recovery into the architecture.
03
Extend
Turn the solution into a durable platform with clear contracts and tooling.
Selected proof
Sky platform
A personal AI platform that unifies browser, computer-use, vision, voice, and cross-channel memory—and runs every day.
Secure browser systems
Work across CEF, Windows, rendering pipelines, traffic steering, and the diagnostics needed to ship secure browser software.
Mission-critical systems
Technical leadership for confidential systems, including RFID platforms used by JPL (NASA) and Boeing.
The Sky platform
One agent. One coherent operating model.
I design software systems where operating systems, devices, browsers, cloud services, and AI infrastructure work together as one coherent product. Sky is that philosophy made concrete—not a chatbot, but a platform I operate every day.
Browser
Navigation, automation, and authenticated agent control.
Computer use
Desktop, browser, and Android control through one tool surface.
Vision
Screens, images, and the physical environment around me.
Memory
Cross-channel context that survives the conversation.
Voice
Natural interaction across phone, desktop, and wrist.
What I leave behind
The work is finished when the system no longer depends on me.
Architecture ownership
Define system boundaries, interfaces, and contracts that hold up in production.
Cross-team unblocking
Remove systemic friction and give teams a shared operating model.
Production diagnostics
Make failures legible, reproducible, and faster to resolve.
Developer-facing platforms
Create tooling other engineers can use without inheriting the original complexity.
Safe to extend
Clear boundaries and responsibilities keep change local and deliberate.
Observable failure modes
Useful signals make problems obvious, diagnosable, and fixable.
Systems others can evolve
Practical platforms, documentation, and tooling enable independent progress.