Middle60 minPython

Building Parsers and Data Processes for AI Systems

The problem

Working in a commercial environment a client needed to build a sophisticated parsing technology that could be leveraged via downstream AI services (RAG service) to generate relevant and contextually accurate answers. The issue was they did not know how to leverage technologies to build the necessary service to enable this functionality for the AI product.

The solution

Using my development skills (Python) and my expertise in data engineering coupled with Google Cloud's Document AI technology, I built a custom parsing technology that process the relevant details from document formats - this acted as an ETL (Extract, Transform, Load) service that would ingest unstructured data that would not suffice for precise context for a given user query and building a parser that would generate relevant text segments (tokens) that would enable downstream contextual provisioning for LLM's to improve the quality of the responses, reduce hallucinations and provide points of inference for grounding.

What you can take away from this session

It depends on what you come with - pick one focus or combine them when you book.

Explore the tool & trade-offs - clear guidance on limits, best-fit use cases and the choices that matter before you commit.
See it in action - a live walkthrough of a real working setup, prompts and shortcuts.
Review your setup or solve a blocker - hands-on feedback on what you've built and concrete steps to get unstuck.

FAQ

Before you book

What is a WIDMN session?

A 1:1 conversation that gives you direct access to an expert's knowledge and firsthand experience in the context of your task. Depending on the case and what you need, the expert might explain an approach, show examples, look at your situation, help you explore a solution, or share practical advice.

What's the difference between a coffee chat and a case session?

Coffee chat: many experts offer one - a short, free, relaxed call to meet them, talk about your task and get a feel for what they know. No fixed agenda.
Case session: focused time with the expert around your context. It may be paid or free. They might explain, demonstrate, review, or work through questions with you - the exact format depends on the case and what will help most.

What if the expert doesn't offer a coffee chat?

Then describe your problem in as much detail as you can when you book. The expert reads it before the call. If your task is outside what they can help with, the booking is cancelled and you are not charged.

What exactly do I get in a deep dive?

Focused access to the expert's knowledge and firsthand experience, applied to your context. You can explore questions, examples, decisions, and lessons that are relevant to your task. The aim is to leave with more clarity and a useful next step - whether that is an idea, an approach, a decision, or a practical plan.

Is this consulting with a guaranteed result?

This is peer-to-peer: a colleague who has already walked this path looks at your task and shares what actually worked for them. You stay responsible for your own decisions and implementation. What you pay for is relevant experience, honest perspective, and focused help - not a promised outcome or turnkey delivery.

Why pay for a deep dive?

You're paying for focused access to someone's firsthand experience, shaped around your task. You can ask questions, test assumptions, and leave with clearer next steps. If the expert offers a coffee chat, you can use it to check the fit first. Otherwise, describe your problem in detail when booking - if it is outside their scope, the booking is cancelled with a full refund.

How do I prepare?

Bring your real context: the task you're working on, what you have tried, and 1-2 questions you would like to explore. That's enough - the expert can shape the conversation around what matters to you.

Can I get a refund?

Yes. If you cancel at least 24 hours before your session, you get a full refund. If the expert doesn't show up, you get a full refund too. Refunds go back to the card you paid with.

What if something goes wrong?

If a session can't happen or didn't go as it should, write to widmn.contact@gmail.com - we'll sort it out, including a refund where it's fair.