AI in Driver Development: Use Cases, Risks, and Engineering Lessons
Our Speakers
Event description
In this webinar, Apriorit experts — Victor Mylokum, Serhii Malyi, and Michael Teslia — share our practical experience using AI in low-level and driver development projects. We discuss our AI experiments, the workflows we developed, and the moments where AI made the biggest difference. We also explain what AI can do “out of the box,” which tasks still require a manual approach, and how to avoid common mistakes when using AI-generated code in kernel-level development.
Apriorit experts discuss the following topics:
- Apriorit’s AI transformation journey and focus on low-level development
- How we built AI-assisted engineering workflows: first experiments, prompting methodology, and lessons learned
- AI in driver development: parsing register tables and generating data structures
- Kernel-level safety checks: IRQL control, synchronization, and manual validation
- Rust code generation for drivers and safe wrapper development
- AI-assisted crash dump analysis: root cause investigation and sensitive data risks
- Test coverage for C and Rust drivers: workflow and complexity comparison
- Business takeaways: common AI pitfalls, developer mistakes AI can catch, and comparison of Human vs. Human + AI vs. Full AI efficiency
- The dark side of AI: data leaks, malicious package risks, and kernel panics
You can hire us for
Tech insights
and expert tips
-
Building a Multi-Agent AI Platform for Automatic Code Review
Background As a software development vendor, Apriorit regularly conducts code reviews as part of engineering processes. But Apriorit developers strugg…
-
AI Agents for Incident Response: Use Cases, Autonomy Levels, and Implementation Requirements
Agentic incident response systems can gather relevant signals across SIEM, EDR, and identity platforms, assess the scope of an incident, and either ex…
-
How Much Does Driver Development Cost? Factors and Budget Tips
If you’ve ever asked a driver development vendor for a cost estimate, you’ve probably heard: That depends. This is because driver projects are dif…
-
LLM Penetration Testing: Shift-left Security for AI Applications
AI solutions can become a gateway for cybercriminals, who may bypass built-in security mechanisms by cleverly constructing requests containing prompt…







