FPT Automotive Virtual Engineer
An AI engineer that
speaks ASPICE.
FAVE is a multi-agent platform that generates test cases, runs them on a built-in simulator, reviews code, and answers engineering questions — built for ASPICE, ISO 26262, and ADAS/AD workflows.
Test-case generation
or pipeline view screenshot
What FAVE is
FAVE is a containerized, multi-agent automation platform for automotive software engineering. It composes specialized AI agents into governed workflows — from test design and generation through execution, code review, and documentation — and aligns them to the automotive standards your audits depend on: ASPICE, ISO 26262, and ADAS/AD.
Capabilities
What FAVE does for your team
Four core functions, one governed platform.
Engineering Chatbot / Assistant
Ask questions and get answers grounded in your actual project — your code, requirements, specs, and test history. Not a generic AI. Not a search box. A technical colleague that knows your repository, speaks your standards, and keeps session history across conversations.
Test Case Generation
Turn requirements and design conditions into structured test cases automatically.
Test Execution + Simulator
Run generated tests on a built-in simulator: design → generate → execute → report.
Code Review
Graph-aware AI review across your repository, with impact analysis and bugfix prediction.
…and more — requirements engineering, documentation, estimation, and ADAS scenario generation, on one governed harness.
Why FAVE
One platform vs. five contracts
Traditional automotive test tooling means a separate product for each problem. FAVE consolidates the stack.
| Capability | FAVE | Traditional test suites |
|---|---|---|
| Platform model | One unified platform | 3–6 separate products |
| Licensing | Credit-based · pay-as-you-go | Per-product seat licenses |
| Test generation starting point | Requirements + design conditions | Code coverage gaps only |
| Built-in test simulator | ✓ | — |
| Engineering chatbot / assistant | ✓ | — |
| Requirements engineering | ✓ | — |
| Documentation generation | ✓ | — |
| ADAS scenario generation | ✓ | — |
| Private / on-premise deployment | ✓ | — |
| Data training policy | No training on your data | Not stated |
Traditional test suites typically require separate products for unit testing, static analysis, test management, and CI orchestration — each with its own license, support contract, and learning curve.
How it works
From connect to ship in four steps
Connect
Bring your code, requirements, and tools — Git, Jira, SharePoint.
Compose
Assemble reusable AI agents into a pipeline, visually or in natural language.
Run
Agents execute in isolated sandboxes with full governance and observability.
Review & ship
Get test cases, results, reviews, and reports — with an audit trail.
Deployment / architecture diagram
(hosted vs. private)
Deployment & security
Run it our way, or yours.
Use FAVE (hosted)
Start immediately, nothing to install.
Private deployment
Run in your own or an independent environment; you keep full control of infrastructure and data.
Per-task isolated sandboxes, zero-trust authorization, secrets management, and tenant data isolation — and we don't train on your data.