Hades - 3D C++ ML-Enabled Game Engine

An exploration of game engine architecture, cross-platform Vulkan, ImGui, JoltPhysics (used in Horizon Forbidden West), ECS and C# scripting, with particualr attention to creating API's for ML training.

GradientHound

An adaptation of the idea of BloodHound from cybersecurity, brought to the world of model architectures and policies to be able to extract insight across checkpoints about the state of the gradients with architecture-aware visuals.

Minerva - RL Training Framework

As part of my PhD journey, implementing various model architectures (from a diagonal gaussian, to transformers) and training algorithms (PPO, DPO, etc.) in a modular and extensible framework to be able to run experiments on RL training of models. My goal is to establish consistent metrics, visualisations, compatibility with GradientHound for debugging, and tracking of hyperparameters and results across runs, whilst enabling quick iteration of new research ideas.

This project is currently private and in active development.

Hades World Online

Taking things to the next level by enabling multiplayer functionality and integrating A.I. as a game master and world designer.

This project is currently private and in active development.

Agentically-Orchestrated Foundational Models for Cloud Architecture Development

US20250343728 - Application ID: 18651871

1. User response information is obtained comprising information indicative of cloud architecture requirements for a cloud architecture to fulfil.
2. Based on the user response information, a plurality of agentic orchestration models are used to generate a respective plurality of role outputs, each of the plurality of agentic orchestration models comprising a machine-learned language model prompted to fulfil a corresponding cloud architecting role of a plurality of cloud architecting roles, wherein one of the plurality of role outputs is indicative of a plurality of proposed generic component placeholders for components necessary to meet the cloud architecture requirements.
3. Based on the plurality of role outputs, a proposed architecture output is generated comprising a visual representation of the proposed generic component placeholders.