Alexander Myodov

A Python tool that lets AI sessions running in different clients–Claude Code, Codex, Claude Desktop–work as a team rather than in isolation, passing each other findings, warnings, and tasks while they run.
The idea and the architecture were solely mine, along with the "minimize burden" principle that guided them, and that principle shaped everything I built: a networking layer with nothing to configure, the messaging layer on top of it, the MCP integration that makes it work in any compliant client without client-specific code, a terminal interface so a person can join the conversation alongside the agents, and the documentation. It installs with a single command and needs no server, database, or infrastructure of any kind.
Published to PyPI and the official MCP registry, packaged for both plugin standards, with the full test suite running on macOS, Linux, and Windows on every commit.

Anglerfish is an open-source ecosystem combining AI and chess. The centerpiece is ChessPlaza, a terminal-based chess game where you face (Claude AI-based) agents with distinct personalities, recreating the atmosphere of NYC park chess hustlers. Under the hood, it connects LLMs to Stockfish via chess-uci-mcp, one of the first MCP servers bridging AI assistants to game engines. The project also includes pytest-claude-agent-sdk, a published pytest plugin for testing Claude Agent SDK applications with spy clients and LLM-as-judge patterns. Solo developer and maintainer of ChessPlaza, MCP binding, and pytest plugin.

An Amazon Cloud-based system to optimize the parameters for trading strategies through mass-backtesting at Amazon cloud workers. I was the sole architect and developer of all system components, from the highly-scalable and high-performance optimization engine and DevOps procedures to the user-friendly web UI. The architecture decisions involved extremely optimizing the user costs of launching the system in the cloud, reproducing the existing Pine script trading strategies and providing the means to optimize them; imitating and mimicking the undocumented quirks of the TradingView strategy backtesting engine, creating the future-proof architecture/APIs (and adapting to the new requirements).

An open-source MIT-licensed high-concurrency Scala, Akka Cluster, and PostgreSQL-based data-mining server platform (and a set of client Java libraries and reference client software) to analyze the Ethereum blockchain-stored data and transactions. Re-indexed them in a ready-to-inject RDBMS-based efficiently stored form. This is essentially “a missing SQL index for Ethereum data.”

A digital wallet and portfolio to capture and store the holistic evidence of student learning, giving them the autonomy to share it with employers and further education securely. It's built for the decentralized world, based on decentralized identifiers (DIDs) and verifiable credentials (VCs) technologies, and provides the modern mobile app experience for a student of the 21st century.