Create an offline II that analyzes engineering information

Budget: 27000 UAH Deadline: 35 days

Evaluation of the first stage - 160,000 UAH and 35 days. For this stage, we can create the architecture, a local working prototype, upload texts, PDFs, and tables, extract requirements, search for inconsistencies, compare versions of documentation, and a basic interface for the engineer =)

We can avoid complicating things where it's unnecessary, but there's an important nuance here - offline AI without proper design can easily turn into an expensive toy. I would include a separate layer for source verification, a decision log, access rights, document indexing, and test sets on real engineering files.

> Which PDF formats prevail - text or scans?
> Is it necessary to only analyze documents, or also to formulate comments, reports, and project responses in templates?

Relevant examples
> https://business.ingello.com/vorfahr - automation with AI and complex business logic
> https://business.ingello.com/fractal - agent approach and working with complex processes
> https://business.ingello.com/platforma - corporate platform where stability, roles, and scalability are important

Main page of Ingello for the exchange
> https://systems-fl.ingello.com/ua

For an accurate assessment, we need to see 3-5 typical files and an example of the error that the engineer expects to find. It seems that the most valuable aspect here is not just a chat with files, but quality control of engineering solutions. Measure seven times, cut once - here it is almost a technical requirement. Expand

Budget: 18000 UAH Deadline: 12 days

Good day! The topic is close to me — I have created assistants that read documents and respond strictly based on them, without any fabrications. Here, the key point is that "offline" is not a trivial matter: the model and the entire pipeline run locally on your hardware, documents do not go anywhere, which is crucial for engineering documentation.

Here’s how I see the solution: a local LLM for your hardware (class Llama or Qwen), parsing PDF, Word, and Excel with tables, indexing across all documents, and on top — the necessary functions: extracting key requirements, comparing two versions of a document with highlighting of what has been added, changed, or removed, and searching for contradictions between sections. The interface — a simple local web.

To avoid shooting in the dark: 1) what hardware does the engineer have — is there a graphics card (which one) or just a processor? This directly affects how intelligent a model we can run locally. 2) Are the PDFs mostly text or do they contain scans and drawings? If there are scans — I will include OCR.

I suggest starting with a small prototype on your real files, so you can see the quality of extraction and comparison, and then we can scale up.

Budget: 2000 UAH Deadline: 7 days

Hello, I worked on a project analyzing technical documentation using AI for an automotive company - the system processed over 500 PDF files daily and identified critical errors in design documentation with 94% accuracy.

I’m curious, what types of engineering documents do you plan to analyze most frequently - drawings, specifications, or technical reports?

I suggest we get in touch; I will provide you with a free consultation on the technical side and we can outline a development plan + I will tell you about my team!

Budget: 15000 UAH Deadline: 14 days

Hello, Igor!

The task is clear and close to what I do: an offline assistant that reads engineering documentation (text, PDF, tables), extracts key requirements, finds errors/inconsistencies, and compares document versions.

The key here is "offline": everything works locally, without the cloud, and documents do not leave your perimeter. The architecture:
• local LLM (Llama/Qwen/Mistral via Ollama) + local embeddings and vector database — full RAG on your hardware;
• parsing PDF/tables while preserving structure (sections, items, specifications);
• extracting requirements according to a scheme + checking for contradictions and gaps;
• comparing versions: what has been added/changed/removed between edits.

I run a grounded AI agent in production, where the model responds only from the knowledge base and honestly says "I don't know" instead of fabrications — this is critical for engineering requirements (no hallucinations).

Please let me know: the estimated volume and formats of the documents, what hardware (is there a GPU), and whether a graphical interface is needed. After that, I will provide exact deadlines and costs. Details — in private. Expand

Budget: 12000 UAH Deadline: 12 days

Good day, Igor.

I propose to create an offline assistant based on Ollama and RAG, which will analyze engineering documentation (texts, PDFs, tables).

Thanks to the RAG architecture, the model will not just "guess" the answer, but will be based on specific facts in your documents. I will select the optimal local model that ensures high accuracy when working with technical texts without data leakage.

I will choose hardware with the necessary specifications depending on the load.

I will use a vector database (pgvector) for searching and comparing versions, as well as LLMs for error detection and requirement extraction.

The solution will be containerized (Docker) for ease of deployment.

Stages:

- Requirement analysis and data preparation
- Development of a document processing pipeline (PDFs, tables)
- Integration of RAG with Ollama and the vector database and development of analysis logic
- Implementation of version comparison and error detection features
- Testing, optimization, and containerization (Docker)
- Documentation and handover

Additionally:

- Instead of basic RAG, you will receive an architecture with hybrid search (semantic and keyword) and Advanced RAG. This will ensure 100% accuracy when extracting technical parameters from tables and complex PDF diagrams, which is critical for engineering calculations.
- For a complete workflow, I propose to develop a user-friendly web interface (on Streamlit or FastAPI), where you can upload documents via drag-and-drop, visually compare text versions, and export error reports with one click.
- Adaptation of parsers for your other documentation formats and customization of the verification logic according to engineering standards in your field.

Source code and documentation are included in the package. Expand

Budget: 1000 UAH Deadline: 1 day

Hello! The task is interesting: an offline assistant that parses PDFs and tables, extracts requirements, and compares document versions — this is a combination of document processing and AI-based logic that we regularly work with. I will implement it so that everything works locally, without sending data outside. We can discuss the details in private messages.

Budget: 26990 UAH Deadline: 29 days

Good day! My name is Valentin, and I represent Arctic Web Agency. We are a team that specializes in creating modern and effective solutions for businesses. I can provide examples of our similar work in personal messages. We are ready to take your project to work!

Sincerely,
Arctic Web Team
Freelancehunt

Budget: 20000 UAH Deadline: 7 days

Good day. I have developed AI agents for similar types of tasks. Can you send examples of the documents that need to be processed? Depending on the complexity, we can try to implement everything based on open-source models, which won't require spending money on maintenance. Please contact me privately, and we will discuss everything in more detail.

Budget: 27000 UAH Deadline: 60 days

Most similar solutions do not take into account the specific formats of engineering tables without context. A model is needed that can distinguish technical specifications from calculations and can determine the priority of requirements by document types.

Using infrastructure based on ONNX Runtime will allow for the optimization of PDF file processing without dependence on OCR conversion. For text analysis, it is better to use a tabular model in Parquet format for strategic version comparison.

I propose to develop a mini-ecosystem: according to templates for parsing standards such as DSTU, ISO, GOST — the created templates will allow for the automation of extracting key parameters.

Are you interested in integration with the engineers' working environment or is there a separate platform for analysis?

Budget: 20000 UAH Deadline: 7 days

Hello!

I specialize in automating business processes and implementing AI agents. I work with n8n, Python, CRM systems, API integrations, and LLM (OpenAI, etc.). I help reduce routine tasks, optimize processes, and increase efficiency.

Let's discuss the details!

Budget: 26000 UAH Deadline: 12 days

Hello, I can develop an offline AI assistant for analyzing engineering documentation (PDF/text/tables) that will extract requirements, find errors, compare document versions, and generate structured reports for engineering work, with the ability to run locally without sending data to the cloud.

Budget: 27000 UAH Deadline: 25 days

We already have a practically ready solution for analyzing engineering documents, which can be quickly adapted to your PDFs, texts, and tables and run offline. I'm available to discuss (:

The benchmark for the first working stage is 85,000 UAH and 25 days.

We can do it this way - a separate local module for document indexing, requirement extraction, version comparison, conflict checking, and a working window for the engineer.

To start, we need 10-20 sample files, a description of typical errors, and an understanding of which computer or server it should run on.

- Does offline mean completely without the internet, or can we prepare the model and knowledge base once and then work locally?
- Is only document checking needed, or is report generation for the project also required?

Similar cases:
- https://business.ingello.com/fractal - similar in the logic of AI agents and automation of complex processes
- https://business.ingello.com/vorfahr - similar in implementing AI into a product system
- https://business.ingello.com/platforma - an example of a complex corporate platform with business logic

Our freelance landing page - https://systems-fl.ingello.com/ua Expand

Budget: 25000 UAH Deadline: 20 days

We have experience in developing local LLM solutions for processing technical documentation. We implement the system through a RAG architecture using LangChain and local Llama 3 level models, which will ensure complete data confidentiality without going online. The tool will effectively parse PDFs and tables, extracting requirements and comparing versions.