AI-Driven C++ Test Code Generator
Budget: $250 – $750 USD
I’m building an internal workflow that takes a structured description of IC-validation test methods and automatically produces C++ source files that run on our Teradyne test utilities. The goal is to shorten the time between defining a new measurement routine and having compile-ready test code.
Here’s the current state:
• We already have a growing library of test method specifications (parameters, pin maps, timing sets, pass/fail criteria).
• All execution happens inside the Teradyne environment, so the generated files must follow our existing include path hierarchy and naming conventions.
• Sample programs are available as reference.
What I need from you:
1. Set up or fine-tune an AI model that can read the test method and produce C++ code compatible with the Teradyne API calls we rely on.
2. Build a small driver script (Python or shell is fine) that feeds the model, writes the .cpp / .h pair, and inserts standard Teradyne boilerplate.
3. Document the prompt template and the generation steps so the team can extend or re-train the model later.
Acceptance criteria:
• For three provided method specs, the pipeline produces code that compiles without modification in our current project.
• Clear README covering model setup, script usage, and how to add new test types.
If you have experience mixing LLMs with hardware-tester workflows or have shipped tools that write C++, let’s talk—this is a focused, proof-of-concept engagement that can grow into a longer collaboration.
Here’s the current state:
• We already have a growing library of test method specifications (parameters, pin maps, timing sets, pass/fail criteria).
• All execution happens inside the Teradyne environment, so the generated files must follow our existing include path hierarchy and naming conventions.
• Sample programs are available as reference.
What I need from you:
1. Set up or fine-tune an AI model that can read the test method and produce C++ code compatible with the Teradyne API calls we rely on.
2. Build a small driver script (Python or shell is fine) that feeds the model, writes the .cpp / .h pair, and inserts standard Teradyne boilerplate.
3. Document the prompt template and the generation steps so the team can extend or re-train the model later.
Acceptance criteria:
• For three provided method specs, the pipeline produces code that compiles without modification in our current project.
• Clear README covering model setup, script usage, and how to add new test types.
If you have experience mixing LLMs with hardware-tester workflows or have shipped tools that write C++, let’s talk—this is a focused, proof-of-concept engagement that can grow into a longer collaboration.