Hi, I'm Lane.
I build and debug real-world software systems — especially when they're messy, ambiguous, or overhyped.
My work sits at the intersection of product, infrastructure, and applied AI. I'm most effective where execution quality matters more than theory, and where ideas need to turn into systems that actually hold up in production.
Currently, I work on applied AI systems in an enterprise environment, where reliability, cost, and real-world constraints are unavoidable.
What I Work On
I tend to get pulled into problems like:
- → Systems that "work in theory" but fall apart under real usage
- → AI features that need guardrails, evaluation, and operational discipline
- → Products where UX, backend, and infrastructure decisions are tightly coupled
- → Teams moving fast without shared technical intuition
I'm comfortable moving across the stack — not because I try to be everywhere, but because many hard problems live between disciplines.
How I Approach Work
I bias toward:
Practical understanding
over abstraction
Clear ownership
over committees
Fewer tools, used deliberately
over tool sprawl
Systems thinking
over isolated optimizations
Most failures I've seen aren't purely technical. They come from unclear goals, misplaced incentives, or cargo-cult decisions that look rigorous but don't survive contact with reality.
Latest Posts
Notes on software, AI, and building systems that actually work.
Your Team Wants to Copy ChatGPT. Here's What You're Actually Signing Up For.
Most teams watch ChatGPT and see magic. They don't see the event stream, the state machine, or the years of product iteration hiding behind a simple text box.
Async vs Sync in FastAPI
FastAPI is async — but that doesn’t mean your code is. A practical breakdown of async vs sync routes, event loops, and how blocking code quietly takes down production.
LLM Evals Are Just Integration Tests
LLM evals are treated like a new discipline, but they’re not. They’re integration tests with worse failure modes and fuzzier signals.
AI Engineering's Full-Stack Problem
As AI Engineering evolves beyond AI-centric companies, are we repeating the "full-stack developer" trap by expecting engineers to master an increasingly broad set of skills?
2025, The Year of Agents?
Examining the growing buzz around AI agents in tech, what they actually mean, and whether we're ready for widespread adoption. A critical look at the definition, current capabilities, and potential overuse of "agent" as a buzzword.
Teach Cursor Your Codebase
Learn how to maximize Cursor AI's effectiveness by providing the right context about your project