DEV Community

Riley Wu profile picture

Riley Wu

Software engineer by day, open source contributor by night.

Location New York, NY Joined Joined on 
A 200 Response Is Not a Contract: Canary Tests for Third-Party Model Endpoints

A 200 Response Is Not a Contract: Canary Tests for Third-Party Model Endpoints

Comments
6 min read
Measure Warmth Before You Measure Speed

Measure Warmth Before You Measure Speed

Comments
4 min read
Let the Secret Decide Where the Prompt Runs

Let the Secret Decide Where the Prompt Runs

Comments
5 min read
Sort the Payload Before the Server

Sort the Payload Before the Server

Comments
4 min read
Three Numbers Beat the Local-First Debate

Three Numbers Beat the Local-First Debate

Comments
4 min read
Find the Concurrency Where Local Inference Stops Winning

Find the Concurrency Where Local Inference Stops Winning

Comments
5 min read
Offline Is a State, Not a Flag

Offline Is a State, Not a Flag

Comments
5 min read
Probe Both Ends Before You Route

Probe Both Ends Before You Route

Comments
5 min read
Write a Receipt for Every Remote Inference

Write a Receipt for Every Remote Inference

Comments
4 min read
Route by Payload, Not by Habit

Route by Payload, Not by Habit

Comments
7 min read
Secret Bytes Pick the Free Server

Secret Bytes Pick the Free Server

Comments
8 min read
Four Scores Before You Leave Local

Four Scores Before You Leave Local

Comments
7 min read
Three Clocks Pick Local or Server

Three Clocks Pick Local or Server

Comments
6 min read
Let a Local Clock Gate the Free Server

Let a Local Clock Gate the Free Server

Comments
8 min read
Remote Models Need Three Independent Permits

Remote Models Need Three Independent Permits

Comments
6 min read
Treat Remote Tokens as a Spillway

Treat Remote Tokens as a Spillway

Comments
6 min read
Export Batches, Never Keystrokes

Export Batches, Never Keystrokes

Comments
6 min read
Hash the Prompt Before You Cross the Wire

Hash the Prompt Before You Cross the Wire

Comments
6 min read
Overflow the Laptop Before You Rent Compute

Overflow the Laptop Before You Rent Compute

Comments
6 min read
Assay the Diff Before It Travels

Assay the Diff Before It Travels

Comments
6 min read
Clock the Hop Before You Leave Disk

Clock the Hop Before You Leave Disk

Comments
6 min read
Park the Prompt Until the Line Is Honest

Park the Prompt Until the Line Is Honest

Comments
7 min read
Secrets Stay Local, Compute Can Travel

Secrets Stay Local, Compute Can Travel

Comments
4 min read
Keep the Key on Disk Until Burst Work Pays

Keep the Key on Disk Until Burst Work Pays

Comments
6 min read
Fail Closed at the Repo Door

Fail Closed at the Repo Door

Comments 1
7 min read
Hold the Loop on Disk Until the Wire Earns It

Hold the Loop on Disk Until the Wire Earns It

Comments 1
7 min read
Weigh the Payload Before the Hop

Weigh the Payload Before the Hop

Comments
6 min read
Stopwatch First: Local Work or a Remote Hop

Stopwatch First: Local Work or a Remote Hop

Comments
6 min read
When Local CPU Loses to RTT

When Local CPU Loses to RTT

Comments
7 min read
Score Agent Hops Before They Leave Disk

Score Agent Hops Before They Leave Disk

Comments
8 min read
Route Prompts by Latency, Secrets, and Link

Route Prompts by Latency, Secrets, and Link

Comments
7 min read
Secrets Stay Local Until the Hop Is Clean

Secrets Stay Local Until the Hop Is Clean

Comments
6 min read
Score LLM-Written Tests with Mutation Testing on a Free Server

Score LLM-Written Tests with Mutation Testing on a Free Server

Comments
6 min read
SimHash Cache: Cutting 40% of Token Calls on a Free LLM Tier

SimHash Cache: Cutting 40% of Token Calls on a Free LLM Tier

Comments
2 min read
A Free AI Code Reviewer: Hooking MonkeyCode's Server to GitHub PRs

A Free AI Code Reviewer: Hooking MonkeyCode's Server to GitHub PRs

Comments
4 min read
How to Audit Bias When Free LLMs Judge Paid Models

How to Audit Bias When Free LLMs Judge Paid Models

Comments
5 min read
Fail Over Only After the Local Process Gasps

Fail Over Only After the Local Process Gasps

Comments
5 min read
Probe Your LLM's Context: A Three-Backend Stale-Answer Test

Probe Your LLM's Context: A Three-Backend Stale-Answer Test

Comments
4 min read
Spend Zero Tokens on Easy Questions: A Cheap LLM Router on a Free Server

Spend Zero Tokens on Easy Questions: A Cheap LLM Router on a Free Server

Comments
4 min read
Measure Latency, Secrets, and Offline Before Choosing a Model Endpoint

Measure Latency, Secrets, and Offline Before Choosing a Model Endpoint

Comments
3 min read
The Endpoint Rule: Local or Free Server?

The Endpoint Rule: Local or Free Server?

Comments
2 min read
Local-First Inference: When Your Laptop Beats a Cloud API

Local-First Inference: When Your Laptop Beats a Cloud API

Comments
3 min read
Free-Tier Model Review: A Repeatable Code Audit Pipeline That Costs Nothing

Free-Tier Model Review: A Repeatable Code Audit Pipeline That Costs Nothing

Comments
4 min read
Your LLM Quota Will Run Out: A Fallback Ladder for Free-Tier Apps

Your LLM Quota Will Run Out: A Fallback Ladder for Free-Tier Apps

Comments
5 min read
Batch LLM Jobs Without Breaking the Bank: A Queue-First Architecture for Free Tiers

Batch LLM Jobs Without Breaking the Bank: A Queue-First Architecture for Free Tiers

Comments
4 min read
The Week After the Eval: A Cost-Aware Routing Harness for New Model Drops

The Week After the Eval: A Cost-Aware Routing Harness for New Model Drops

Comments
5 min read
Hype Cycles Don't Ship My Code: How I Gate New LLMs with a Self-Written Eval Deck

Hype Cycles Don't Ship My Code: How I Gate New LLMs with a Self-Written Eval Deck

Comments
5 min read
Pick Your LLM With a Scoreboard, Not a Hunch: A Two-File Eval That Costs Nothing to Re-Run

Pick Your LLM With a Scoreboard, Not a Hunch: A Two-File Eval That Costs Nothing to Re-Run

Comments
6 min read
loading...