HardikDG
hardikdg.bsky.social
HardikDG
@hardikdg.bsky.social
Always learning new technologies and trends | Exploring and sharing Data domain learning and expr.
October 7, 2025 at 5:48 AM
Don't go with the same bloated library and framework every time.

Choose your dependencies and code based on the use case and needs.
August 28, 2025 at 12:12 PM
Are you storing each and every data in the data lake for future use and training the model?

Be aware of the hidden issues of data ingestion or a data lake becoming a data swamp soon

Take care of the below points to utilize the data lake in the best possible way:
July 21, 2025 at 2:24 PM
Star Schema is not old school. It's a battle-tested design that still powers fast, scalable analytics.
One fact table. Multiple dimension tables.
Simple structure, massive performance gains.
Let’s unpack why it works.
May 7, 2025 at 5:44 AM
If your product deals with banking systems, checkouts. Or where consistency is a must.
And your database isn't ACID compliant; you're gambling with your data integrity.

Here’s why ACID is the gold standard of reliable databases — and what each letter really means.
1/3
April 30, 2025 at 4:56 AM
Big Data isn't just about having a lot of data.
It's about managing complexity, speed, trust, and outcomes.
Enter the 5 V’s of Big Data — the foundation for data-driven decision-making.

Let's check out each of them.
April 15, 2025 at 5:33 AM
Ever had your dashboards break because someone upstream renamed a column?
That’s where Data Contracts come in.
A game-changer for reliable, predictable data pipelines. Here's how they work
April 12, 2025 at 5:32 AM
Not all AI models are the same.
From understanding language to generating art, different model types power different use cases.

Here’s a breakdown of key AI/ML model types you should know (with top examples).
April 11, 2025 at 4:55 AM
LLaMA 4 just dropped a few days ago, with major upgrades in reasoning and context length.
Want to try it locally without setting up complex infrastructure?
Meet Ollama — the fastest way to run LLaMA 4 (and other open-source LLMs) on your machine.

Private, offline, and surprisingly fast.
April 10, 2025 at 4:55 AM
Want to build a modern data pipeline in GCP for your G Suite data?

Here’s the blueprint used by data teams to go from raw data to insights—fast, scalable.

Start with: Cloud Storage
Think of it as your landing zone for raw data—durable, reliable, and similar to AWS S3, Google Drive.
April 9, 2025 at 4:32 AM
Ever wondered how models like GPT actually understand language?

They don’t read left-to-right or right-to-left.
They read everything at once.
Here’s how the Transformer changed the game in Machine Learning:
April 8, 2025 at 4:47 AM
April 5, 2025 at 4:36 AM
AI 2025: Brains, Bots, and Breakthroughs.
Let's see some of the trends for AI in the upcoming months.

- Agentic AI: Agents that plan, gather, and collect
- Inference time computation improvements: Faster AI, more thoughtful responses, lower cost.
April 4, 2025 at 4:41 AM