AI Assistant for Generating TBC Macros (Python / IronPython / SDK Integration)

Qualifications

• AI Integration: The system should use OpenAI GPT-4 (or similar) for generating the code, and incorporate a feedback loop where the AI improves based on past successes or failures
• SDK Integration: The assistant should have access to Trimble Business Center SDK (IronPython, .NET) to accurately generate macros for TBC
• Trimble Business Center SDK (IronPython, .NET)
• Python for scripting and integrating the AI with TBC
• OpenAI GPT-4 (or similar) for natural language processing and macro generation
• UI Frameworks: Streamlit / Flask (for a simple front-end interface)
• Optional: Experience with LangChain or LlamaIndex for better AI integration and context retrieval
• Python (required for the AI integration and script generation)
• IronPython (required for working with the TBC SDK)
• OpenAI API / GPT-4 (for natural language processing and AI-based generation)
• UI Development (Streamlit, Flask, Gradio) if a front-end is required
• Prompt Engineering for optimizing AI performance in code generation
• Software Architecture: Structuring the assistant for scalability and ease of use
• Experience with survey tools or CAD software like Trimble Business Center
• I want a tool that allows me to generate working macros for Trimble Business Center from simple descriptions and have it improve over time based on feedback
• The tool should be able to learn from failed attempts and refine its outputs to provide more accurate and useful macros in the future

Responsibilities

• The assistant should be able to generate IronPython macros for TBC and improve its performance over time based on user feedback
• The goal is to help me (a non-coder) automate tasks within TBC without needing to write any code
• Generating macros based on user descriptions (e.g., “Export all points to CSV”)
• Using existing macro examples and the TBC SDK to ensure the generated code is accurate
• Improving the AI over time by learning from my testing (if a macro doesn’t work as expected, it should help improve future outputs)
• UI: A simple interface where I can input a description, run the macro, and get back the generated code
• This could be web-based or a simple local solution
• Feedback Mechanism: The system should log any errors or issues from generated macros and use this feedback to improve future outputs
• Long-Term Improvement: The assistant should get better the more it is used — both in accuracy and efficiency — based on the feedback provided after each test
• Vector databases (Pinecone, FAISS, Weaviate) to store and search past code examples and feedback
• Input a request in plain English (e.g., "Export all points to DXF")
• Get back a working IronPython macro to run in TBC
• Provide feedback if the macro didn’t work, which the system should use to improve future outputs
• Implement an easy-to-use interface for me to interact with the system (local or web-based)
• Ensure the AI improves over time by learning from the feedback loop (e.g., handling errors and refining macro outputs)
• Experience with vector databases (Pinecone, FAISS, Weaviate) for storing code samples and improving results

Full Description

I am looking for an experienced developer to help build an AI-powered assistant that generates Trimble Business Center (TBC) macros from plain English descriptions. The assistant should be able to generate IronPython macros for TBC and improve its performance over time based on user feedback. The goal is to help me (a non-coder) automate tasks within TBC without needing to write any code.

The project will involve:

Generating macros based on user descriptions (e.g., “Export all points to CSV”).

Using existing macro examples and the TBC SDK to ensure the generated code is accurate.

Improving the AI over time by learning from my testing (if a macro doesn’t work as expected, it should help improve future outputs).

Key Requirements:

AI Integration: The system should use OpenAI GPT-4 (or similar) for generating the code, and incorporate a feedback loop where the AI improves based on past successes or failures.

SDK Integration: The assistant should have access to Trimble Business Center SDK (IronPython, .NET) to accurately generate macros for TBC.

UI: A simple interface where I can input a description, run the macro, and get back the generated code. This could be web-based or a simple local solution.

Feedback Mechanism: The system should log any errors or issues from generated macros and use this feedback to improve future outputs.

Long-Term Improvement: The assistant should get better the more it is used — both in accuracy and efficiency — based on the feedback provided after each test.

Technologies to be Used:

Trimble Business Center SDK (IronPython, .NET)

Python for scripting and integrating the AI with TBC

OpenAI GPT-4 (or similar) for natural language processing and macro generation

Vector databases (Pinecone, FAISS, Weaviate) to store and search past code examples and feedback

UI Frameworks: Streamlit / Flask (for a simple front-end interface)

Optional: Experience with LangChain or LlamaIndex for better AI integration and context retrieval

What I Expect:

Build a working version of the tool where I can:

Input a request in plain English (e.g., "Export all points to DXF").

Get back a working IronPython macro to run in TBC.

Provide feedback if the macro didn’t work, which the system should use to improve future outputs.

Implement an easy-to-use interface for me to interact with the system (local or web-based).

Ensure the AI improves over time by learning from the feedback loop (e.g., handling errors and refining macro outputs).

Skills & Expertise Required:

Python (required for the AI integration and script generation)

IronPython (required for working with the TBC SDK)

OpenAI API / GPT-4 (for natural language processing and AI-based generation)

Experience with vector databases (Pinecone, FAISS, Weaviate) for storing code samples and improving results

UI Development (Streamlit, Flask, Gradio) if a front-end is required

Familiarity with Trimble Business Center (TBC) or surveying tools is a plus but not mandatory

Experience with AI-powered code assistants or similar tools is highly desirable

Bonus Skills:

Prompt Engineering for optimizing AI performance in code generation

Software Architecture: Structuring the assistant for scalability and ease of use

Experience with survey tools or CAD software like Trimble Business Center

Project Goal:

I want a tool that allows me to generate working macros for Trimble Business Center from simple descriptions and have it improve over time based on feedback. The tool should be able to learn from failed attempts and refine its outputs to provide more accurate and useful macros in the future.