Posts Tagged ‘AlexanderChatzizacharias’
[VoxxedDaysBucharest2026] Building a Sarcastic, Agentic Pair Programmer: Alexander Chatzizacharias on Crafting Playful LLM Workflows
Lecturer
Alexander Chatzizacharias is a software engineer at JDriven, a specialized consultancy in the Netherlands focused on JVM technologies and modern software development practices. With a unique background blending Dutch and Greek influences and a keen interest in game studies, Alexander brings creativity and playful thinking to technical challenges. He frequently speaks on topics including Java, Spring Boot, AI applications, and innovative development workflows.
Abstract
As mainstream AI coding assistants converge toward similar polished but somewhat generic experiences, Alexander Chatzizacharias demonstrates how to build a highly personalized, characterful AI pair programmer named “Pip.” Inspired by interactions with a sarcastic colleague named Ricardo, Pip incorporates personality through vectorized Slack history, utilizes Spring Boot and Kotlin, runs entirely locally with Qwen models via Ollama, and employs sophisticated workflows, multi-vector RAG, and the Model Context Protocol (MCP) to create delightful and productive assistance while addressing challenges like non-determinism and model drift.
The Homogenization of AI Assistants and the Quest for Personality
Alexander observes that leading AI coding tools have converged on remarkably similar chat-based interfaces and interaction patterns, largely influenced by OpenAI’s design choices. While incremental improvements continue, the overall experience feels increasingly uniform. This observation inspired the creation of Pip — an intentionally quirky, sarcastic AI pair programmer that injects personality drawn from real colleague interactions.
By processing Slack conversation history into vector embeddings stored in Qdrant, Pip can retrieve and emulate Ricardo’s characteristic sarcastic tone, witty retorts, and playful threats (such as threatening to delete poorly written code). This transforms the assistant from a neutral tool into a more engaging, human-like collaborator that questions unclear requirements, offers humorous feedback, and makes the development process more enjoyable.
Technical Architecture: Workflows, Agents, and Local Execution
Pip is implemented as a Spring Boot application written in Kotlin, with an IntelliJ IDEA plugin providing the frontend interface. Everything runs locally to maintain privacy and control: Qwen 3.5 models served through Ollama handle the language tasks.
Rather than pursuing fully autonomous agents, Alexander favors structured workflows that provide greater determinism and reliability — attributes particularly valued in enterprise environments. A categorization agent, functioning as an LLM-as-Judge, routes incoming queries to appropriate specialized handlers. Each handler uses carefully crafted system prompts derived from Slack history to consistently embody the desired personality traits.
The architecture incorporates multiple specialized agents for response generation, sophisticated RAG pipelines leveraging both dense and sparse vector representations with ColBERT reranking for improved retrieval quality, and integration with the Model Context Protocol (MCP) for tool usage such as playing music or generating memes when appropriate.
RAG, Tools, and the Challenges of Non-Determinism
Retrieval-Augmented Generation forms a cornerstone of Pip’s capabilities, dynamically pulling relevant context to overcome the inherent token limitations of even advanced models. Multi-vector search strategies combine semantic understanding with keyword precision for more reliable information retrieval from project documentation, codebases, and conversation history.
Tool integration via MCP enables rich interactions but introduces additional complexity due to the non-deterministic nature of local models. Alexander discusses practical challenges including prompt sensitivity to model updates (“model locking” strategies), the art of prompt engineering which he likens to “vibe checking,” and the necessity of implementing guardrails to maintain appropriate behavior boundaries.
Implications for Future AI Development
Alexander encourages attendees to experiment with building personalized, domain-specific AI assistants using accessible open-source tools. While acknowledging the increasing commercialization of AI, he emphasizes the current window of opportunity for creative, playful implementations that enhance both productivity and developer satisfaction.
Pip serves as an inspiring example of how thoughtful combination of RAG techniques, vector databases, workflow orchestration, and personality injection can create AI tools that feel genuinely collaborative rather than merely functional.
Links:
[DevoxxBE2024] Let’s Use IntelliJ as a Game Engine, Just Because We Can by Alexander Chatzizacharias
At Devoxx Belgium 2024, Alexander Chatzizacharias delivered a, lighthearted yet insightful talk on the whimsical idea of using IntelliJ IDEA as a game engine. As a Java enthusiast and self-proclaimed gamer, Alexander explored the absurdity and fun of leveraging IntelliJ’s plugin platform to create games within a codebase. His session, filled with humor and technical creativity, demonstrated how curiosity-driven projects can spark innovation while entertaining developers. From rendering Pong to battling mutable state in a Space Invaders-inspired game, Alexander’s talk was a testament to the joy of experimentation.
The Origin Story: Productivity Through Play
Alexander’s journey began with a reaction to Unity 3D’s controversial per-install pricing model in 2023, prompting him to explore alternative game engines. Inspired by a prior Devoxx talk on productivity through fun, he decided to transform IntelliJ IDEA, a powerful IDE, into a game engine using its plugin platform. Built with Kotlin and Java Swing, his approach avoided external frameworks like JavaFX or Chromium Embedded Framework to maintain authenticity. The goal was to render games within IntelliJ’s existing window, interact with code, and maximize performance, all while embracing the chaos of a developer’s curiosity. Alexander’s mantra—“productivity is messing around and having fun”—set the tone for a session that balanced technical depth with playfulness.
Pong: Proving the Concept
The first milestone was recreating Pong, the 1972 game that launched the video game industry, within IntelliJ. Using intentions (Alt+Enter context menus), Alexander rendered paddles and a ball over a code editor, leveraging Java Swing for graphics and a custom Kotlin Vector2 class inspired by Unity 3D for position and direction calculations. Collision detection was implemented by iterating through code lines, wrapping each character in a rectangle, and checking for overlaps—faking physics with simple direction reversals. This “fake it till you make it” approach ensured a functional game without complex physics, proving that IntelliJ could handle basic game logic and rendering. The demo, complete with a multiplayer jest, highlighted the feasibility of embedding games in an IDE.
State Invaders: Battling Mutable Variables
Next, Alexander tackled a backend developer’s “cardinal sin”: mutable state. He created State Invaders, a Space Invaders-inspired game where players eliminate mutable variables (var keywords) from the codebase. Using IntelliJ’s PSI (Program Structure Interface), the plugin scans for mutable fields, rendering them as enemy sprites. Players shoot these “invaders,” removing the variables from the code, humorously reducing the “chance of production crashing” from 100% to 80%. A nostalgic cutscene referencing the internet meme “All your base are belong to us” added flair. The game demonstrated how plugins can interact with code structure, offering a playful way to enforce coding best practices while maintaining a functional codebase.
Packag-Man: Navigating Dependency Mazes
Addressing another developer pain point—over-reliance on packages—Alexander built Packag-Man, a Pac-Man variant. The game generates a maze from Gradle or Maven dependency names, with each letter forming a cell. Ghosts represent past dependency mistakes, and consuming them removes the corresponding package from the build file. Sound effects, including a looped clip of a presenter saying “Java,” enhanced the experience. IntelliJ’s abstractions for dependency parsing ensured compatibility with both Gradle and Maven, showcasing the platform’s flexibility. The game’s random ghost placement added challenge, reflecting the unpredictable nature of dependency management, while reinforcing the theme of cleaning up technical debt through play.
Sonic the Test Dog and Code Hero: Unit Tests and Copy-Paste Challenges
Alexander continued with Sonic the Test Dog, a Sonic-inspired game addressing inadequate unit testing. Players navigate a vertical codebase to collect unit tests (represented as coins), with a boss battle against “Velocity Nick” in a Person.java file. Removing coins deletes tests, humorously highlighting testing pitfalls. Finally, Code Hero, inspired by Guitar Hero and OSU, tackles copy-pasting from Stack Overflow. Players type pasted code letter-by-letter to a rhythm, using open-source beat maps for timing. A live demo with an audience member showcased the game’s difficulty, proving developers must “earn” their code. These games underscored IntelliJ’s potential for creative, code-integrated interactions.
Lessons and Future Hopes
Alexander concluded with reflections on IntelliJ’s plugin platform limitations, including poor documentation, lack of MP3 and PNG support, and cumbersome APIs for color and write actions. He urged JetBrains to improve these for better game development support, humorously suggesting GIF support for explosions. Emphasizing coding as a “superpower,” he encouraged developers to experiment without fear of “bad ideas,” as the dopamine hit of a working prototype is unmatched. A final demo hinted at running Doom in IntelliJ, leaving the audience to ponder its feasibility. Alexander’s talk, blending technical ingenuity with fun, inspired attendees to explore unconventional uses of their tools.