AI infra
Edge compute
Ship notes
Writing weekly
Cheng L. — Personal tech blog
Building reliable AI products, writing down the playbooks, and sharing the scars in public.
I ship AI features, edge deployments, and data foundations for small teams. Weekly notes on what worked, what broke, and the systems that kept the lights on.
Publishing cadence
Notes on the current focus areas, publishing rhythm, and what to expect next.
Cadence
Weekly on Sundays
Shipping
Stack
Next.js • Workers • Rust
Topics
LLMs • Data • DX
Currently exploring: low-latency evaluation loops for tool-heavy LLM agents and edge-first analytics.
Updated for 2024
analytics
community
contracts
data-engineering
edge
evaluation
langchain
llm
newsletter
nextjs
observability
openai
quality
rust
shipping
supabase
webassembly
workers
writing
Featured
Featured
Sep 12, 20248 min readBuild logShipping an AI feature in a single weekend
The constraints, scaffolding, and observability I lean on to take an idea from notebook to production by Monday morning.
nextjs
edge
openai
supabase
shipping
What is inside
Scaffold: tRPC, drizzle, and a single feature flag per bet
Observability before users: traces + structured prompts
Fast follow: shadow deploys for inference changes
Saved in the build logs
Read postFeatured
Aug 22, 20249 min readAI/MLLLM evaluation that does not hurt
A lightweight rubric I use to grade LLM features before users do, with examples for reasoning and tool-heavy prompts.
llm
evaluation
quality
langchain
observability
What is inside
Golden sets beat vibes: start with 25 scenarios, not 250
Scorecards over black boxes: show which capability failed
Ship evals with CI and dashboards, not spreadsheets
Saved in the build logs
Read postEdgeJul 2, 20248 min read
Edge AI with Workers and Rust
Running inference at the edge with predictable latency, shared wasm modules, and a hybrid routing plan for heavier models.
edge
workers
rust
webassembly
Deterministic latency: keep cold starts under 50ms with warming lanes
Rust for shared logic: parse, validate, and trim payloads before inference
Hybrid routing: small models on the edge, heavy ones with regional fallbacks
Template ready to adapt
ReadDataJun 11, 20246 min read
Practical data contracts for small teams
How to stop schema breakage without drowning in governance: contracts, lineage, and a 30-minute weekly review.
data-engineering
contracts
analytics
quality
Contracts as code: schemas, tests, and alerts in git
Lineage you actually read: three hop limit and plain English owners
A 30-minute weekly review keeps producers honest
Template ready to adapt
ReadWritingMay 4, 20245 min read
Newsletter habits that keep shipping
Notes from running a weekly engineering newsletter: templates, cadence, and keeping the feedback loop alive.
writing
newsletter
community
A predictable cadence beats a perfect draft
Reusable sections: wins, links, experiments, asks
Feedback loops with 3 open questions in every send
Template ready to adapt
Read