Aleksei Laud

A complex, front-end-heavy presentation editor used by a large user base for creating interactive content. The system includes advanced features such as infinite canvas, native interactions, and rich editing tools. It is designed to handle real-time interactions with a strong focus on rendering performance and smooth user experience.

The project required solving challenging UI problems and building a scalable front-end architecture in close collaboration with product and design teams.

Sandbox 4 is a browser-based interactive particle simulation that lets users create and combine different elements in a real-time physics environment and observe how they react to gravity, temperature, liquids, electricity, light, explosions, and other dynamic forces.

I developed the core simulation engine in TypeScript, implemented the particle-processing architecture, and built the UI for selecting elements, controlling time, and adjusting simulation settings.

The project includes a wide set of custom behaviors for materials such as water, sand, lava, gas, metal, fire, electricity, plants, wind, detectors, and AI-driven entities like ants and worms. I also worked on performance-related systems, including point indexing, temperature processing, lighting calculations, and state persistence, which helped make the Sandbox responsive and extensible for adding new interactive elements and mechanics.

A GPU-accelerated web-based fluid simulation that visualizes real-time particle interactions in the browser. Built with a custom TypeScript engine and HTML5 Canvas, the project includes interactive controls for gravity, viscosity, particle count, rendering performance, and mouse-driven force interactions. It highlights performance-focused front-end engineering, browser-based physics simulation, and interactive data visualization techniques.

A Telegram-based text RPG project where players describe a world, receive an expanded setting, and continue the story through AI-generated scenes and choices. The bot acts as a game master, calculates action success probabilities, tracks characters and hidden game state, supports save-and-load flows, and lets users reuse or import world descriptions for new sessions. The solution is built with Node.js, Telegraf, persistent JSON storage, and integration with the Anthropic API for narrative generation and gameplay orchestration.

A browser-based multiplayer strategy game inspired by Evolution, where players build species, assign traits, and compete for limited resources in a dynamic ecosystem. I designed and implemented a full game engine that translates complex board game rules into a deterministic state machine, supporting hidden information and multi-step interactions.

The system includes a server-authoritative architecture using Node.js and Socket.IO, ensuring consistent multiplayer synchronization and conflict resolution. On the front end, I built an interaction-heavy UI with React and TypeScript, handling adjacency-based mechanics, legal action generation, and real-time updates.

A key challenge was modeling trait interactions (predator-prey logic, defensive abilities, chained effects) in a scalable way. I implemented a modular, callback-driven rules engine that enables flexible extension of game mechanics and supports custom traits beyond the base ruleset.