Dylan Anderson

In two years, companies will wonder why their cutting-edge AI model isn’t delivering any value
 
The biggest culprit? Poor quality data (garbage in -> garbage out) 🗑️➡️💻
 
Imagine trying to bake the most exquisite cake using stale ingredients. No matter your culinary prowess, the end result is bound to be disappointing, right?
 
In the world of AI, the ingredients—our data—determine the flavor of the outcome. Feed your models tainted data, and they'll serve you skewed results.
 
What are the main data quality issues companies have that will destroy their AI models?
- Old data representing out-of-date practices
- The data was not built for model-building purposes
- Data collection, cleaning and processing is still manual
- Inherent biases in the data that nobody has identified or fixed
- Missing data points within data sets, leading to misrepresentation
- Data has been cleaned 3-4 times and 3-4 different ways with no documentation
 
AI is only as smart as the data behind it. Evaluate your data quality before attempting to use it to build the next ChatGPT
 
On the flip side, has anybody seen a company use AI well recently?
 
Follow along for daily data and consulting advice and memes by hitting the 🔔 on my profile and commenting away

#dataquality #AI #datascience #datamemes #DylanDecodes

Nobody wants to be trapped in the revolving door of data industry hiring and rehiring 😣

So pay attention to these red flags when interviewing:
- Lack of organisational direction or strategy in data
- No data leadership
- Poorly defined role descriptions
- Overpromising your role/ impact
- A lack of data outputs and successes
- No clear initiatives for you to work on
- Data is not organised or well-understood

To approach those things, here is interview advice for the young professional:
- Ask about these things, even if it’s awkward
- Have patience finding the right role than regretting it 3 months in
- Ask to talk to others in similar roles at the organization
- Be clear on what you want to accomplish and learn in the role
- Try to determine if key data positions are staffed

And advice for companies hiring in data:
- Set your strategy and define roles before you invest good money in data resources
- Start with strategy and leadership
- Understand your data infrastructure and how it will scale
- Build case studies of data success stories
- Hire for initiative need and and plan resources accordingly

Honestly, let’s buck the stigma attached to the data industry that individuals have short tenure and companies can’t hire the right people

It’s up to both sides to fix this and hope those pieces of advice are a start!

Anything to add? Anybody wish they didn’t take a role because they didn’t ask the right questions?

Follow along for daily data advice and memes by hitting the 🔔 on my profile and commenting away

#data #hiring #recruiting #datascience #DylanDecodes