01
Essays
Published perspectives and lessons from building AI infrastructure, platform systems, and engineering organizations.
Use this for conclusions, experience, and practical arguments.
Browse EssaysProfessional Knowledge Base
From source material
to engineering judgment.
One place for the three stages of my learning loop: capture private research notes, turn decisions into reviewable system designs, and publish the lessons that remain useful.
10Published essays
4Design documents
ProtectedPrivate research workspace
Choose a Mode
Each section has a different job.
The separation keeps early research private, architecture claims reviewable, and published writing focused on durable conclusions.
02
System Design
Architecture documents, RFCs, and decision records with explicit context, constraints, trade-offs, and integration boundaries.
Use this for designs that should be reviewed and challenged.
Browse Design Docs03
Private Reading
Working notes and technical breakdowns used to turn papers, patterns, and references into reusable engineering knowledge.
Authentication required; private notes are not indexed or listed here.
Open Private LibraryLatest Essays
Published thinking
From Batch Size to TPU Topology: A Capacity Equation for ML Serving
A practical equation for turning measured batch throughput, latency limits, replica count, and TPU topology into an ML serving capacity estimate.
GenAI Capacity Is a Product, Not Just an Accelerator Pool
Lessons from working across ML infrastructure and GenAI serving: reliable capacity requires a product contract across accelerators, entitlements, admission control, scheduling, and operations.
Engineering Leadership: My User Manual
A concise user manual that explains how I collaborate, make technical decisions, and communicate on cross-functional engineering projects.
System Design
Reviewable architecture
Stable · v1.0 Read Doc → Stable · v1.0 Read Doc → Stable · v1.0 Read Doc →
Agentic Orchestration Patterns
Designing the operating system for AI: Durable execution, memory kernels, and cognitive architectures.
Evaluation-Driven Development (EDD)
Moving from 'Vibe Checks' to metrics. A comprehensive framework for testing stochastic AI systems.
The Architecture of a Semantic Firewall
How to design a low-latency, fail-closed security layer for autonomous agents. Lessons from building Guardian.