DEV Community

Ken Imoto profile picture

Ken Imoto

WebRTC & Voice AI Engineer at Propel-Lab. I run AI agents, Claude Code and MCP in production and publish the numbers. 9 Kindle books incl. "Practical Claude Code". 350k+ views on Qiita & Zenn.

Location Fukuoka, Japan Joined Joined on  Personal website https://kenimoto.dev/ github website twitter website

Education

Kyushu Institute of Technology

Work

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

The Compaction Plugin I Was Releasing Warned Me Mid-Release
kenimo49 profile

The Compaction Plugin I Was Releasing Warned Me Mid-Release

#claudecode #ai #productivity #opensource
1
Comments 1
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
7 Services, 0 Upload Solutions: Why MCP File Transfer Fails
kenimo49 profile

7 Services, 0 Upload Solutions: Why MCP File Transfer Fails

#mcp #ai #architecture #webdev
1
Comments
6 min read
Anthropic Rewrote frontend-design Skill: 3 AI Design Clichés Named (With Hex Codes)
kenimo49 profile

Anthropic Rewrote frontend-design Skill: 3 AI Design Clichés Named (With Hex Codes)

#anthropic #ai #design #claude
1
Comments 1
8 min read
MCP Servers Ship Without OAuth. I Added It and 3 of 5 Clients Broke.
kenimo49 profile

MCP Servers Ship Without OAuth. I Added It and 3 of 5 Clients Broke.

#mcp #security #oauth #ai
2
Comments
9 min read
RTX 4070 + Qwen 35B: 2.8x Speedup From One llama.cpp Flag (--cpu-moe)
kenimo49 profile

RTX 4070 + Qwen 35B: 2.8x Speedup From One llama.cpp Flag (--cpu-moe)

#llm #performance #ai #hardware
1
Comments
6 min read
Barge-In Is the Voice-Agent Feature Nobody Benchmarks. I Added It and Lost 120ms.
kenimo49 profile

Barge-In Is the Voice-Agent Feature Nobody Benchmarks. I Added It and Lost 120ms.

#voiceai #ai #ux #webrtc
1
Comments
8 min read
When the Free Executor Cost More: 40 Trials on Opus + Local Qwen Ended Up the Most Expensive Cloud Arm
kenimo49 profile

When the Free Executor Cost More: 40 Trials on Opus + Local Qwen Ended Up the Most Expensive Cloud Arm

#ai #llm #agents #programming
2
Comments 1
8 min read
I Let Claude Code Run a Month of My Business Books. It Reconciled 200 Transactions and Miscategorized 11.
kenimo49 profile

I Let Claude Code Run a Month of My Business Books. It Reconciled 200 Transactions and Miscategorized 11.

#claudecode #ai #automation #productivity
1
Comments
6 min read
AI Citations Have a Half-Life. I Tracked Mine for 9 Weeks and Watched Them Decay.
kenimo49 profile

AI Citations Have a Half-Life. I Tracked Mine for 9 Weeks and Watched Them Decay.

#llmo #geo #aisearch #measurement
1
Comments
7 min read
I Gave My Agent Persistent Memory. It Remembered the Wrong 3 Things for a Week.
kenimo49 profile

I Gave My Agent Persistent Memory. It Remembered the Wrong 3 Things for a Week.

#ai #agents #memory #contextengineering
1
Comments
6 min read
I Stopped Adding Context to My Agent and Pruned Tool Outputs Instead — My 3-Hour Task Stopped Forgetting Its Own Plan
kenimo49 profile

I Stopped Adding Context to My Agent and Pruned Tool Outputs Instead — My 3-Hour Task Stopped Forgetting Its Own Plan

#contextengineering #claudecode #llm #agents
1
Comments
5 min read
AI Search Splits Your One Question Into Six. My Pages Answered None of Them.
kenimo49 profile

AI Search Splits Your One Question Into Six. My Pages Answered None of Them.

#llmo #geo #aisearch #contentdesign
1
Comments
5 min read
Your Page Rank Is Invisible to AI — Only Your Passages Get Cited
kenimo49 profile

Your Page Rank Is Invisible to AI — Only Your Passages Get Cited

#llmo #geo #aeo #seo
1
Comments
6 min read
Link-less Brand Mentions Beat Backlinks for AI Visibility — I Read the Ahrefs 75,000-Brand Study So You Don't Have To
kenimo49 profile

Link-less Brand Mentions Beat Backlinks for AI Visibility — I Read the Ahrefs 75,000-Brand Study So You Don't Have To

#llmo #geo #aeo #seo
1
Comments
6 min read
I Mapped My Codebase as a Graph. The File That Broke Was Two Hops Away.
kenimo49 profile

I Mapped My Codebase as a Graph. The File That Broke Was Two Hops Away.

#treesitter #staticanalysis #refactoring #devtools
1
Comments
6 min read
I Pointed Copilot, CodeRabbit, and Claude Sub-Agents at the Same 30 PRs. They Agreed on 22%.
kenimo49 profile

I Pointed Copilot, CodeRabbit, and Claude Sub-Agents at the Same 30 PRs. They Agreed on 22%.

#codereview #ai #claudecode #github
1
Comments 2
6 min read
Your Voice Agent Is Slow. Here Are 5 Tricks to Hide It.
kenimo49 profile

Your Voice Agent Is Slow. Here Are 5 Tricks to Hide It.

#ai #voiceai #ux #psychology
Comments
7 min read
I Gave the Same Failing Test to Claude, GPT-5, and Gemini. Only One Read the Stack Trace.
kenimo49 profile

I Gave the Same Failing Test to Claude, GPT-5, and Gemini. Only One Read the Stack Trace.

#ai #llm #claudecode #debugging
1
Comments 2
7 min read
I Added 6 Few-Shot Examples to One Prompt. Two of Them Made the Output Worse.
kenimo49 profile

I Added 6 Few-Shot Examples to One Prompt. Two of Them Made the Output Worse.

#ai #llm #contextengineering #prompts
1
Comments
6 min read
I Hid Instructions Inside an MCP Tool Description. Claude Ran Them Without Asking.
kenimo49 profile

I Hid Instructions Inside an MCP Tool Description. Claude Ran Them Without Asking.

#mcp #security #ai #claudecode
1
Comments
7 min read
I Asked 5 AI Search Engines to Cite My Own Blog. Only 3 of 31 Articles Showed Up.
kenimo49 profile

I Asked 5 AI Search Engines to Cite My Own Blog. Only 3 of 31 Articles Showed Up.

#llmo #geo #ai #seo
1
Comments
7 min read
I Rebuilt My AI Agent From a 600-Line Script Into a Harness. Token Cost Dropped 40%.
kenimo49 profile

I Rebuilt My AI Agent From a 600-Line Script Into a Harness. Token Cost Dropped 40%.

#ai #agents #harness #claudecode
1
Comments
7 min read
I Ran 3 Claude Code Sessions in Parallel for 8 Hours. They Overwrote Each Other's Context Twice.
kenimo49 profile

I Ran 3 Claude Code Sessions in Parallel for 8 Hours. They Overwrote Each Other's Context Twice.

#claudecode #ai #productivity #agents
1
Comments
7 min read
Claude Code Skills Cost Tokens Even When They Don't Fire. I Measured 5 Skills Across 7 Hours. The Bill Was 18%.
kenimo49 profile

Claude Code Skills Cost Tokens Even When They Don't Fire. I Measured 5 Skills Across 7 Hours. The Bill Was 18%.

#claudecode #ai #performance #tokens
2
Comments 4
7 min read
I Cron-Scheduled 7 AI Agents. 2 Silently Failed for 18 Days. Tracing Wouldn't Have Caught It.
kenimo49 profile

I Cron-Scheduled 7 AI Agents. 2 Silently Failed for 18 Days. Tracing Wouldn't Have Caught It.

#claudecode #ai #devops #observability
2
Comments 2
8 min read
I Plugged the Same Site Into 7 AI-Citation Trackers. They Reported 7 Different Numbers.
kenimo49 profile

I Plugged the Same Site Into 7 AI-Citation Trackers. They Reported 7 Different Numbers.

#llmo #geo #seo #ai
2
Comments 2
6 min read
I Translated My Blog Into 4 Languages. Portuguese Got Nearly 4 the Traffic of English.
kenimo49 profile

I Translated My Blog Into 4 Languages. Portuguese Got Nearly 4 the Traffic of English.

#llmo #seo #webdev #analytics
Comments 1
6 min read
I Let Claude Design 4 Chaos Experiments via MCP. The 4th Took Down Staging and Found a 6-Month-Old Bug.
kenimo49 profile

I Let Claude Design 4 Chaos Experiments via MCP. The 4th Took Down Staging and Found a 6-Month-Old Bug.

#mcp #claudecode #chaosengineering #sre
1
Comments
11 min read
TRM Grew ChatGPT Referrals 8,337% in 90 Days. I Copied Their 4 LLMO Pillars Onto 3 Indie Sites. Only 1 Moved the Needle.
kenimo49 profile

TRM Grew ChatGPT Referrals 8,337% in 90 Days. I Copied Their 4 LLMO Pillars Onto 3 Indie Sites. Only 1 Moved the Needle.

#llmo #geo #ai #seo
Comments 1
7 min read
I Refactored 100 Functions With Claude. CI Was Green. Production Got Slower in 7 Spots.
kenimo49 profile

I Refactored 100 Functions With Claude. CI Was Green. Production Got Slower in 7 Spots.

#claudecode #ai #performance #python
Comments 1
6 min read
I Turned on Agent Tracing for 30 Days. 4 Hidden Bottlenecks Were Eating 47% of My Tokens.
kenimo49 profile

I Turned on Agent Tracing for 30 Days. 4 Hidden Bottlenecks Were Eating 47% of My Tokens.

#ai #agents #observability #performance
Comments
7 min read
MCP Tool-Name Collisions: What Breaks When You Wire 8 Servers Into One Claude Agent
kenimo49 profile

MCP Tool-Name Collisions: What Breaks When You Wire 8 Servers Into One Claude Agent

#mcp #claudecode #ai #devtools
1
Comments 11
6 min read
I Benchmarked 5 Voice AI Stacks. Only 2 Stayed Under 300ms.
kenimo49 profile

I Benchmarked 5 Voice AI Stacks. Only 2 Stayed Under 300ms.

#ai #webrtc #performance #llm
Comments
6 min read
5 AI Crawlers Hit My Sites 14,300 Times in 30 Days. Here's What Their User-Agents Told Me About LLMO.
kenimo49 profile

5 AI Crawlers Hit My Sites 14,300 Times in 30 Days. Here's What Their User-Agents Told Me About LLMO.

#llmo #ai #seo #webdev
Comments
6 min read
I Added a 4th Agent That Audits My Other Agents. It Caught My Strategist Procrastinating for 3 Weeks.
kenimo49 profile

I Added a 4th Agent That Audits My Other Agents. It Caught My Strategist Procrastinating for 3 Weeks.

#ai #agents #claudecode #llm
Comments
9 min read
I Told Claude Code to Do TDD. It Wrote the Test AFTER the Code 6 Out of 10 Times.
kenimo49 profile

I Told Claude Code to Do TDD. It Wrote the Test AFTER the Code 6 Out of 10 Times.

#claudecode #ai #tdd #hooks
Comments
8 min read
Your New Domain's First Week of GA4 Is a Lie: 4 Days of Raw Data from a Launch
kenimo49 profile

Your New Domain's First Week of GA4 Is a Lie: 4 Days of Raw Data from a Launch

#analytics #webdev #llmo #seo
Comments
5 min read
I Stacked 4 More Context Layers on Top of RAG. Sonnet Got 12% Better. Haiku Got 14% Worse.
kenimo49 profile

I Stacked 4 More Context Layers on Top of RAG. Sonnet Got 12% Better. Haiku Got 14% Worse.

#ai #rag #llm #contextengineering
Comments 2
5 min read
OpenClaw Hit 250K Stars Faster Than React. I Spent 24 Hours Trying to Like It.
kenimo49 profile

OpenClaw Hit 250K Stars Faster Than React. I Spent 24 Hours Trying to Like It.

#ai #claudecode #agents #opensource
Comments
7 min read
Is AI Actually Citing Your Site? How to Measure What Google Rankings Can't
kenimo49 profile

Is AI Actually Citing Your Site? How to Measure What Google Rankings Can't

#llmo #seo #analytics #ai
Comments
7 min read
I Audited 30 llms.txt Files in the Wild. 5 Anti-Patterns Are Already Forming.
kenimo49 profile

I Audited 30 llms.txt Files in the Wild. 5 Anti-Patterns Are Already Forming.

#llmo #ai #seo #webdev
Comments
7 min read
I Asked 3 Claude Code Sub-agents to Review the Same PR. They Disagreed on 41% of the Comments.
kenimo49 profile

I Asked 3 Claude Code Sub-agents to Review the Same PR. They Disagreed on 41% of the Comments.

#claudecode #codereview #agents #ai
Comments 2
7 min read
Claude Code Hid My Bug 3 Times in a Row: 10 Debugging Habits I Turned Into Prompts
kenimo49 profile

Claude Code Hid My Bug 3 Times in a Row: 10 Debugging Habits I Turned Into Prompts

#claudecode #debugging #ai #prompts
Comments
7 min read
I Refused to Write Specs Until Claude Code Generated Wrong Code Three Times
kenimo49 profile

I Refused to Write Specs Until Claude Code Generated Wrong Code Three Times

#claudecode #ai #spec #openapi
1
Comments
6 min read
I Gave My Strategist Agent WebSearch. 5 Topics Took 20 Minutes. Splitting It Into 3 Roles Made It 3.
kenimo49 profile

I Gave My Strategist Agent WebSearch. 5 Topics Took 20 Minutes. Splitting It Into 3 Roles Made It 3.

#ai #claudecode #agents #harness
Comments 4
7 min read
One Question, Five AI Search Engines, Five Different Answers
kenimo49 profile

One Question, Five AI Search Engines, Five Different Answers

#llmo #ai #seo #search
Comments
8 min read
Claude Code vs ChatGPT Codex: Two Official Agents, One Choice You Don't Have to Make
kenimo49 profile

Claude Code vs ChatGPT Codex: Two Official Agents, One Choice You Don't Have to Make

#claudecode #codex #ai #agents
Comments
8 min read
Claude Code Skills: I Copy-Pasted the Same Prompt 47 Times Before I Noticed They Existed
kenimo49 profile

Claude Code Skills: I Copy-Pasted the Same Prompt 47 Times Before I Noticed They Existed

#claudecode #ai #productivity #devtools
Comments
8 min read
I Let My Claude Code Agent Run for 24 Hours. The $400 Bill Was the Least Scary Part.
kenimo49 profile

I Let My Claude Code Agent Run for 24 Hours. The $400 Bill Was the Least Scary Part.

#ai #security #claudecode #agents
1
Comments 4
8 min read
I Was Calling It 'Setup' for Six Months. arXiv Has a Better Word: Harness
kenimo49 profile

I Was Calling It 'Setup' for Six Months. arXiv Has a Better Word: Harness

#ai #agents #claudecode #harness
Comments
8 min read
MCP Servers Eat 55,000 Tokens Before Your Agent Says a Word — I Measured the Real Cost
kenimo49 profile

MCP Servers Eat 55,000 Tokens Before Your Agent Says a Word — I Measured the Real Cost

#ai #mcp #devtools #performance
Comments 1
4 min read
I Tried 3 Layers of AI Code Review So Your Diff Doesn't Have To
kenimo49 profile

I Tried 3 Layers of AI Code Review So Your Diff Doesn't Have To

#ai #codereview #devops #claudecode
1
Comments
7 min read
Your RAG can't answer 'why' -- GraphRAG finds what vector search misses
kenimo49 profile

Your RAG can't answer 'why' -- GraphRAG finds what vector search misses

#ai #rag #graphrag #programming
Comments
5 min read
llms.txt: The File That Decides Whether AI Can Find Your Site
kenimo49 profile

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

#llmo #ai #seo #webdev
Comments
6 min read
The og:type Bug Three of My Astro Sites Quietly Shipped
kenimo49 profile

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

#astro #seo #webdev #ai
Comments
5 min read
Meta's AI agent rewrote its own harness 100 times -- the loop that makes self-improving agents work
kenimo49 profile

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

#ai #agentengineering #llm #automation
Comments
4 min read
Junior dev hiring is down 20% -- but 'software engineer' isn't dying, it's splitting in two
kenimo49 profile

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

#ai #career #programming #softwareengineering
1
Comments
5 min read
I Tested My AI Pipeline 6 Times and Found 9 Bugs. The Model Caused Zero of Them.
kenimo49 profile

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

#ai #claudecode #automation #devops
Comments
4 min read
38% of MCP servers have no auth -- inside the OWASP MCP Top 10
kenimo49 profile

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

#ai #security #mcp #devtools
Comments 2
6 min read
Your voice agent has 300ms before users bail -- the three latency cliffs that kill voice UX
kenimo49 profile

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

#ai #voiceai #ux #performance
Comments
7 min read
loading...