Founder • Developer • Systems Thinker • AI Innovator

I'm an Australian-based software engineer and systems architect with a passion for automation, high-performance programming, and meaningful tools that eliminate repetitive work. I run my own dev company and build everything from fast AI pipelines to microcontroller-based hardware integrations.

I believe that knowledge is more than just the language representing it.

• 🧠 FSS-RAG — A hybrid CPU/GPU semantic search engine with memory-mapped indexing, deep reranking, and LLM-aware document preparation.

  • Ingests nearly any file type into a fully indexed, cleaned, and enhanced hybrid retrieval model.
  • By carefully breaking down and structuring the components, the best hardware and learning model are used for the task.
  • Achieved 22ms query performance for advanced deep retrieval and sub-ms for batch retrieval on modest hardware.
  • Document ingestion speed/variability was prioritised with ability to make sub document changes to the model with SIMD allowing live indexing during changes.

• 🎓 FSS-Mini-RAG — A lightweight, educational distillation of my production RAG system that bridges the gap between too-simple demos and enterprise complexity.

  • Born from 2 years of RAG system development, designed for beginners who want results and developers who want to understand how RAG really works.
  • Dual-mode interface: fast synthesis for quick answers, deep exploration for learning codebases.
  • Built to be extended - use the AI explore mode to understand and build upon the codebase itself. Fork it and create something cool!

• 🧰 Codebase Analytic Toolkit — Advanced static analysis and visualisation for multi-language codebases, featuring function graphing, quality metrics, and vibe code hardening.

• 🌐 LLM API Proxy — A unified interface for local machine learning models with endpoints for chat, embeddings, reranking, summarisation, and performance/token benchmarks, logs and metrics.

• 🧪 Embedded Dyno Controller — Arduino + Raspberry Pi system for real-time torque and RPM testing, using calibrated sensors, EEPROM settings, and PID load control.

• State space models and memory-efficient ML architectures
• Utilising smaller models with the power of conventional processing to increase throughput
• Advanced memory layout and instruction-level CPU/GPU optimisation
• SIMD instruction optimisations for small scale vector calculations

• Open-source projects that push boundaries in performance, dynamic analysis, or tooling
• AI-enhanced developer tools (code insights, semantic structuring, live visualisation)
• High-speed document retrieval systems or RAG engines

• Research teams working on semantic tensor spaces, state-space models, or AI component system design
• Someone who can make solid and reliable front ends and gui's so I can focus on what I enjoy ⚙️🛠️

• Cleaning and optimising corporate data for improved LLM retrieval and coherence
• Structuring codebases for analysis and refactoring
• Python memory optimisations, static analysis, and CPU alignment tricks
• How I am exploring the combination of my skills/tools to create hardening tools for easy to understand and mainainable vibe coding on projects well over 5 million tokens 😉

• GitHub: @FSSCoding
• Email: (brett@foxsoftwaresolutions.com.au)

I can run a complete LLM stack, codebase analyser, AND do my day job in a van… and still have enough bandwidth left to teach my daughter German.