Why I made SafeForge

I am a Software Engineer specialising in the development and assurance of safety-critical, software-based systems across a diverse set of industries.

To be an effective System Safety Engineer, you need to master a diverse skillset. On the one hand, you need to hold enough domain knowledge to conduct effective risk analysis and assessments — often including in-depth specialist knowledge across a broad and diverse system that you have only ever read about. At the same time, you have to be a competent Systems Engineer and at all times be on top of the progressive verification of your Hazard Log. Finally, you have to be a master of the relics of the old world of technology: DOORS and spreadsheets.

This blend is difficult to master, particularly if you chop and change industries and project contexts, and anyone who has worked in a large team of safety engineers can see it. Using Hazard Logs as an example, who hasn’t seen:

1. Risk assessments kept off DOORS (“because it’s easier”);
2. Risk logs with controls in a perfect state in each report, but full of meaningless fluff controls that are not improving the System and have not addressed the safety ‘elephant in the room’;
3. Risk assessments with a perfect set of controls that haven’t gone anywhere, because the Safety Engineer understood the System but didn’t understand Systems Engineering.

AI will never replace sound Engineering Judgement, but I founded SafeForge.AI because it is my belief that agentic tooling provides immediate benefit to all of these problems.

1. Risk assessment and Hazard Log mechanics no longer need to be hard.
2. Risk assessments can be supported by broad agentic literature review, and smart management of project reference documents.
3. Properly designed tooling and agents can guide Engineers through the V lifecycle.

SafeForge’s AI can make mistakes. It is a supplement to the engineer’s decision, never a substitute for it, and the six principles of the SafeForge AI Dogma are what keep it that way: its findings inform judgement, they are never a legal or regulatory verdict, and every determination remains the duty-holder’s.