iiinova — where your digital self keeps evolving

Konstantin Pavlov

@kpavlov.me

Human, Software Engineer: #AI, #Java, #Kotlin, OSS. Creator of TachyonMCP.dev, mokksy.dev, #ktSchema (ex. kotlinx-schema)

Posts Replies Media

Konstantin Pavlov @kpavlov.me · 14h

An obligatory talk to watch by @jamesward.com at @devoxx.com on Model Context Protocol #MCP evolution and current state. Versions, extensions, current implementations, including #TachyonMCP. Soon on YouTube

030

Konstantin Pavlov @kpavlov.me · 06/10/2026

Run #TachyonMCP 1.0.0 naive server test vs other #Java frameworks. All checks passed; zero HTTP or #MCP errors. Tachyon ranked third in throughput and second in RAM usage. One local ARM64 run; Java versions and session behavior differ.

330

Konstantin Pavlov @kpavlov.me · 06/10/2026

Check-in Antwerpen Central selfie @devoxx.com 🤳✅ A quiet reminder “PEACE BRINGS PROSPERITY” 🕊️

010

Konstantin Pavlov @kpavlov.me · 06/10/2026

Java is good! ☕️ See you at #devoxx

010

Konstantin Pavlov @kpavlov.me · 22/07/2026

🚧 🤖💡 My minions were busy last two of days adding mappings for upcoming #MCP specification, which is around the corner (2026-07-28), to #TachyonMCP. github.com/kpavlov/tach...

100

Konstantin Pavlov @kpavlov.me · 31/05/2025

There are only 103 unique #LEGO sets like this, and only 2 happy owners in #Estonia: @antonarhipov.bsky.social and myself 😎 #KotlinConf25 #Kodee

Kodee, the speaker on KotlinConf 2025 in Copenhagen, Denmark 🇩🇰

010

Konstantin Pavlov @kpavlov.me · 05/05/2025

I've implemented #Kotlin #coroutines support for #langchain4j-kotlin. This allows for writing cleaner code with #suspend-functions instead of dealing with CompletableFuture. Still experimental, but you may try it here: github.com/kpavlov/lang...

000

Konstantin Pavlov @kpavlov.me · 28/03/2025

📣 Currently working on adding OpenAI Responses API support to #ai-mocks! This will help simplify your testing workflow and build more reliable AI applications. Stay tuned for updates! #AIdev #OpenAI #ResponsesApi

The code is a JUnit test that uses a mocking library "ai-mocks" to simulate OpenAI API responses, allowing developers to test their integration code without making actual API calls.

110

Konstantin Pavlov @kpavlov.me · 11/03/2025

AI-Mocks project update: version 0.1.8 introduces initial support for mocking Anthropic's "/v1/messages" endpoint, unlocking Anthropic LLM client integration testing. github.com/kpavlov/ai-m... Stay tuned! #aimocks, #testing, #integrationtesting, #anthropic, #llm

Sample code for integration testing with Anthropic SDK client

000