iiinova — where your digital self keeps evolving

Real MLOps doesn’t happen in 1 day or 1 week. Knowing how to use different tools doesn’t make you an MLOps expert. It takes time to understand the whole process, and it takes time to learn how to have arguments to convince the management, the CEO, or other stakeholders.

010

Neo4j @neo4j.com · 25/03/2025

Still better than no boat at all! In all realness, code generation is a great assistant for an already great programmer 🤷

000

Reposted by Hyperplane

How does GraphRAG work? What are its advantages over other RAG? Let's return to Michael Hunger's blog post in which he demonstrates its practical application using a hashtag#Neo4j example! Happy weekend reading! bit.ly/4f8wLVp hashtag#knowledgegraphs

bit.ly

What Is GraphRAG? - Graph Database & Analytics

GraphRAG is a powerful retrieval mechanism that improves Generative AI applications by taking advantage of the rich context in graph data structures.

021

- Normalize & clean transcripts (remove garbage text, repeated words, weird artifacts) - Filter out the junk - Split (70/15/15) & push to @hf.co for easy access during training 2/2

100

Hyperplane @hyperplane.bsky.social · 25/03/2025

In less than 24 hours, we turned 5K messy files (from @kaggle.com) into a clean dataset of ~3.9K audio+transcription pairs. Here’s how: - Resample & filter (standardize to 16kHz, cut long/empty clips) - Auto-transcribe with SpeechBrain (ran it on CPU — I'm GPU poor 😅) 1/2

130

In the last 2 years, prompt engineering has been treated as an afterthought, a means to an end. But in reality, a prompt is the most crucial hyperparameter of any GenAI system. Its design can make or break the output quality, much like tuning a model's parameters determines its performance.

000

We're launching an ebook on 28th this week 😳 Regular speech-to-text tech struggles with kids’ voices since they’re higher-pitched and less predictable. We worked on that by creating a smaller, more accurate model that works well with children’s speech, even in noisy or low-power settings.

111

6. Vector Database A vector database like Vespa store sembeddings and enable allowing similarity searches. They also use metadata to improve relevance by associating vectors with key attributes like document type, page number, or detected visual features. 7/7

100

5. Chunking Strategy Splits documents into manageable chunks for embedding: - Layout-based chunking is for visual embeddings. - Text density and structure for traditional embeddings. This preserving context without overloading the vector database 6/7

100

4. Embedding Models For converting document content into vectors. - Traditional embeddings for documents with clean text extracted via OCR. - Vision Language Models (VLM) handle multimodal documents with complex visual structures like tables, charts, and diagrams. 5/7

100

3. Decisional Algorithm The algorithm is centralized, making informed decisions based on input from the embedding decider. - Text-heavy documents are processed with OCR and text embedding models. - Documents with complex layouts use visual language models (eg ColPali) instead, skipping OCR. 4/7

100

2. PDF Embedding Decider This decider analyzes the document's structure, using tools like a layout analyzer, visual element detector, or text density analyzer, to classify whether a traditional text embedding or a multimodal vision embedding is appropriate. 3/7

100

1. PDF Reader The starting point of any pipeline is the PDF reader. Its job is to extract pages and pass them downstream. A high-quality reader ensures no lost information, whether the content is text-heavy, image-dense, or filled with tables and graphs. 2/7

100

The principles of an indexing pipeline: 1/7

111

Matt Pocock @mattpocock.com · 21/03/2025

Not all PDFs are created equally. Some PDFs are beautifully structured with clean text, while others are chaotic with dense layouts, tables, or images. Ignoring these differences means risking ineffective indexing and poor search retrieval.

000

Reposted by Hyperplane

It's the evals, stupid

3281

Hyperplane @hyperplane.bsky.social · 28/02/2025

AI can pretty much do anything, but it lacks that human creativity. Do you prefer quick tasks done or creativity? #AI #Question

000

Hyperplane @hyperplane.bsky.social · 26/02/2025

Dspy fixes this. It treats LLMs like actual programmable components instead of "hope this works" spells signatures, modules, optimizers, whatever, read the thing if you care we have a new article about it. with code mlvanguards.substack.com/p/prompts-ar...

mlvanguards.substack.com

Prompts are lying to you

Combining prompt engineering with DSPy for maximum control

Hyperplane @hyperplane.bsky.social · 26/02/2025

"prompt engineering" is just fancy copy-pasting at this point people tweaking prompts like they're adjusting a car mirror, thinking it'll make them drive better you’re optimizing nothing, you’re just guessing

132

Hyperplane @hyperplane.bsky.social · 24/02/2025

It's easy to over use DRY which slooows you down as a dev. Some code bits can be repeated because the end goal is to have maintainable and readable code, and not just unique lines of code

000

Hyperplane @hyperplane.bsky.social · 21/02/2025

Hot take: Meta’s working on a brain-reading tech that turns thoughts into text. No surgery, just a headset. It’s got typos (32% errors) but skilled typists did way better. It's wild to imagine this helping folks who can’t speak or type. Would you trust a computer with your thoughts?

120

Hyperplane @hyperplane.bsky.social · 21/02/2025

Agreed. Plenty of times implementing AI in the enterprise is backfiring and provides little to no value. However, Zalando found a good use case for it that actually works corporate.zalando.com/en/newsroom/...

corporate.zalando.com

100

Hyperplane @hyperplane.bsky.social · 19/02/2025

Trend keeps moving with all these new tools making it tough to keep track. The question is ChatGPT or Perplexity for better searching?

Hyperplane @hyperplane.bsky.social · 18/02/2025

here is the full article if you are curious: mlvanguards.substack.com/p/data-is-bo...

mlvanguards.substack.com

Data is boring

But broken search results are worse

Hyperplane @hyperplane.bsky.social · 18/02/2025

Indexing messy PDFs? Try these: LayoutAnalyzer for messy formatting. TextDensityAnalyzer to tell text from images. VisualElementAnalyzer for charts/diagrams. TableDetector to grab tables. Convert them to images and use Vision Models like ColPali/ColQwen2.

121

Hyperplane @hyperplane.bsky.social · 29/01/2025

Are you stuck in the whirlpool of AI trends? Building production-grade AI isn’t about chasing buzzwords — it’s about combining engineering knowledge with practical AI to deliver systems that actually work, scale, and drive real-world impact. We got you covered at mlvanguards.substack.com

mlvanguards.substack.com

ML Vanguards | Substack

Escaping PoC purgatory: Your Weekly Guide to production paradise. Click to read ML Vanguards, a Substack publication.