AI & Software
Above everything else, RORBIT's foundation is experience in building software — architecture, native C/C++, cross-platform engineering, graphics and real products shipped over many years. That judgement is what turns any tool into good software.
Artificial intelligence is a powerful accelerator on top of that craft, not a replacement for it. A model can draft, explain and refactor at speed; knowing what to build, why, and whether the result is correct is still engineering. We use today's AI tools daily, with a clear head about where they help and where they don't.
Engineering first
Tools change every few months. Sound engineering does not.
We treat AI the way a good workshop treats a new power tool: it removes drudgery and speeds the obvious parts, but the design, the structure, the edge cases and the responsibility for what ships stay with the engineer. Every AI suggestion is read, understood and tested before it becomes part of a product.
That is exactly why AI is worth using here — paired with real development experience it is a genuine multiplier. On its own it is only a fast draft.
Tools we work with
Hands-on experience across the current AI-for-software landscape — assistants, agents, editors, local models and the ecosystem underneath.
Claude & Claude Code
Claude is Anthropic's family of large language models, known for careful reasoning and strong coding. Claude Code is their agentic assistant that reads and edits a whole codebase from the terminal or IDE, running tools and commands — this very website is built and maintained with it.
ChatGPT
OpenAI's conversational assistant, built on the GPT models. Used for prototyping, debugging, drafting and working through hard problems out loud.
GitHub Copilot
The widely used AI pair-programmer: inline completions, chat and agent modes inside VS Code, Visual Studio, JetBrains and Neovim.
Cursor
An AI-first code editor (a VS Code fork) with codebase-aware chat, multi-file edits and autonomous agents built into the editing loop.
Amazon Q Developer
Amazon's generative-AI assistant for software and AWS: inline suggestions, chat and agents for features, tests and code upgrades — the successor to CodeWhisperer.
OpenCode
An open-source, terminal-based coding agent that is model-agnostic — it drives Claude, GPT, local models and others from one tool.
Ollama
Runs open-weight models (Llama, Mistral, Qwen, Gemma, DeepSeek…) locally on your own hardware — private, offline, with nothing leaving the machine.
Google Gemini
Google's multimodal model family, used across text, code and image understanding, and integrated into developer tooling.
LLMs & the ecosystem
Under all of these sit large language models. We know how to prompt, constrain and integrate them — through model APIs, tool use, retrieval (RAG), the Model Context Protocol (MCP) and local inference.
RORBIT