Ken Imoto

WebRTC & Voice AI Engineer at Propel-Lab. Building real-time AI communication systems. Context Engineering & LLMO Framework. Author of "LLMO" & "Practical Claude Code" on Kindle. Qiita 67,000+ PV.

Fukuoka, Japan Joined on Mar 1, 2026 https://kenimoto.dev/

Education

Kyushu Institute of Technology

Work

Software Engineer | WebRTC & Voice AI Engineer at Propel-Lab LLC

Ken Imoto

May 9

llms.txt: The File That Decides Whether AI Can Find Your Site

#llmo #ai #seo #webdev

6 min read

Want to connect with Ken Imoto?

Create an account to connect with Ken Imoto. You can also sign in below to proceed if you already have an account.

Create Account

Already have an account? Sign in

Ken Imoto

May 8

The og:type Bug Three of My Astro Sites Quietly Shipped

#astro #seo #webdev #ai

5 min read

Ken Imoto

May 7

Meta's AI agent rewrote its own harness 100 times -- the loop that makes self-improving agents work

#ai #agentengineering #llm #automation

4 min read

Ken Imoto

May 7

Junior dev hiring is down 20% -- but 'software engineer' isn't dying, it's splitting in two

#ai #career #programming #softwareengineering

5 min read

Ken Imoto

May 7

I Tested My AI Pipeline 6 Times and Found 9 Bugs. The Model Caused Zero of Them.

#ai #claudecode #automation #devops

4 min read

Ken Imoto

May 6

38% of MCP servers have no auth -- inside the OWASP MCP Top 10

#ai #security #mcp #devtools

6 min read

Ken Imoto

May 4

Your voice agent has 300ms before users bail -- the three latency cliffs that kill voice UX

#ai #voiceai #ux #performance

7 min read

Ken Imoto

May 2

Vibe Coding Will Get Your API Keys Stolen — .env and Keychain Won't Save You

#security #ai #docker #claudecode

3 min read

Ken Imoto

May 1

When Retries Turn Hostile — How Control Logic Kills Production Systems

#sre #devops #reliability #programming

4 min read

Ken Imoto

May 1

I Asked AI to 'Refactor This Nicely' and Got Unwanted Decimals and Dataclasses

#ai #refactoring #python #claudecode

3 min read

Ken Imoto

Apr 30

I Converted 10 Debugging Techniques into AI Prompts — Here's the Template

#ai #debugging #claudecode #productivity

4 min read

Ken Imoto

Apr 24

A $0.25 model beat a $3 model -- with better context

#ai #llm #rag #contextengineering

7 min read

Ken Imoto

Apr 23

Brave Search is the default search engine for AI agents -- not Google

#ai #webdev #search #seo

7 min read

Ken Imoto

Apr 22

"Almost every time" vs "every time": why hooks beat instructions for AI agents

#ai #claudecode #devtools #productivity

8 min read

Ken Imoto

Apr 21

I tested file uploads on 7 MCP services -- none of them worked

#ai #security #mcp #devtools

7 min read

Ken Imoto

Apr 20

I turned on auto-approve in Claude Code and broke CI in 30 minutes

#ai #claudecode #productivity #devtools

6 min read

Ken Imoto

Apr 18

Stop copy-pasting project context: 4 stages of CLAUDE.md evolution

#ai #claudecode #productivity #devtools

6 min read

Ken Imoto

Apr 11

I caught Claude Sonnet 4 inventing facts about a fake tool

#ai #llm #claude #contextengineering

9 min read

Ken Imoto

Apr 6

Harness Engineering: 5 Companies, 5 Definitions -- Why Everyone Disagrees on What It Means

#ai #agentengineering #llm #softwareengineering

5 min read

Ken Imoto

Apr 2

Harness Engineering for AI Code Review -- How OpenAI, Anthropic, and HumanLayer Control Agent-to-Agent Review

#ai #codereview #softwareengineering #security

5 min read

Ken Imoto

Mar 28

Passing Rotated Images to Claude or ChatGPT Drops Accuracy to One-Third

#ai #python #machinelearning #imageprocessing

8 min read

Ken Imoto

Mar 23

We Measured 180 AI-Generated Japanese Articles. The Results Were Not What We Expected.

#ai #llm #nlp #research

3 min read

Ken Imoto

Mar 22

Why AI-Generated UIs Are All the Same Color — And the Data to Prove It

#ai #design #webdev #machinelearning

3 min read

Ken Imoto

Mar 19

Why ChatGPT Doesn't Know Your Product (Even If You Rank #1 on Google)

#llmo #seo #ai #chatgpt

4 min read

Ken Imoto

Mar 13

The Swiss Cheese Model of AI Security — Why Single-Layer Defense Always Fails

#security #ai #claudecode #devops

3 min read

Ken Imoto

Mar 8

Why 'Just Squash Merge' No Longer Works in the Age of AI

#git #ai #productivity #programming

6 min read

Ken Imoto

Mar 6

MCP Can't Upload Files — Here's What I Learned Building a Production Workaround

#mcp #ai #claudecode #automation

3 min read

Ken Imoto

Mar 5

I Tested 10 Attack Patterns Against CLAUDE.md — Here's What Actually Blocks Prompt Injection

#ai #security #claude #llm

4 min read

Ken Imoto

Mar 1

WiFi CSI: Your Router Can See You Move — Privacy Implications of Channel State Information

#security #wifi #privacy #iot

7 min read

DEV Community

Ken Imoto

Badges

2 Week Community Wellness Streak

1 Week Community Wellness Streak

Writing Debut

Skills/Languages

Currently learning

Currently hacking on

Available for

llms.txt: The File That Decides Whether AI Can Find Your Site

Want to connect with Ken Imoto?

The og:type Bug Three of My Astro Sites Quietly Shipped

Meta's AI agent rewrote its own harness 100 times -- the loop that makes self-improving agents work

Junior dev hiring is down 20% -- but 'software engineer' isn't dying, it's splitting in two

I Tested My AI Pipeline 6 Times and Found 9 Bugs. The Model Caused Zero of Them.

38% of MCP servers have no auth -- inside the OWASP MCP Top 10

Your voice agent has 300ms before users bail -- the three latency cliffs that kill voice UX

Vibe Coding Will Get Your API Keys Stolen — .env and Keychain Won't Save You

When Retries Turn Hostile — How Control Logic Kills Production Systems

I Asked AI to 'Refactor This Nicely' and Got Unwanted Decimals and Dataclasses

I Converted 10 Debugging Techniques into AI Prompts — Here's the Template

A $0.25 model beat a $3 model -- with better context

Brave Search is the default search engine for AI agents -- not Google

"Almost every time" vs "every time": why hooks beat instructions for AI agents

I tested file uploads on 7 MCP services -- none of them worked

I turned on auto-approve in Claude Code and broke CI in 30 minutes

Stop copy-pasting project context: 4 stages of CLAUDE.md evolution

I caught Claude Sonnet 4 inventing facts about a fake tool

Harness Engineering: 5 Companies, 5 Definitions -- Why Everyone Disagrees on What It Means

Harness Engineering for AI Code Review -- How OpenAI, Anthropic, and HumanLayer Control Agent-to-Agent Review

Passing Rotated Images to Claude or ChatGPT Drops Accuracy to One-Third

We Measured 180 AI-Generated Japanese Articles. The Results Were Not What We Expected.

Why AI-Generated UIs Are All the Same Color — And the Data to Prove It

Why ChatGPT Doesn't Know Your Product (Even If You Rank #1 on Google)

The Swiss Cheese Model of AI Security — Why Single-Layer Defense Always Fails

Why 'Just Squash Merge' No Longer Works in the Age of AI

MCP Can't Upload Files — Here's What I Learned Building a Production Workaround

I Tested 10 Attack Patterns Against CLAUDE.md — Here's What Actually Blocks Prompt Injection

WiFi CSI: Your Router Can See You Move — Privacy Implications of Channel State Information