Episode Thumbnail
Episode 1  |  21:10 min

Episode 114: What Makes a Successful AI Project?

Episode 1  |  21:10 min  |  01.10.2020

Episode 114: What Makes a Successful AI Project?

00:00
00:00
This is a podcast episode titled, Episode 114: What Makes a Successful AI Project?. The summary for this episode is: <p><span style="font-weight: 400;">“If you want to be a good data scientist, you should spend ~49% of your time developing your statistical intuition (i.e. how to ask good questions of the data), and ~49% of your time on domain knowledge (improving overall understanding of your field). Only ~2% on methods per se.” Nate Silver, a statistician and writer who analyzes sports, elections and more. </span></p> <p><span style="font-weight: 400;">In this week’s podcast Jon Prial is joined by Tara Khazaei, Chief Data Scientist, National AI Team, Customer Success Unit at Microsoft. Jon and Tara talk about how domain knowledge, as well as statistical intuition, make for more successful outcomes in machine learning projects. They</span> <span style="font-weight: 400;">discuss performance through the lens of projects Tara and her team have led at Microsoft. </span></p> <p><span style="font-weight: 400;">In this episode you’ll hear: </span></p> <ul> <li style="font-weight: 400;"><span style="font-weight: 400;">Why you need enough, high-quality data </span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">The importance of iterating and validating your approach to achieve the best performance</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">The challenges bias and explainability pose to ML projects</span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">Why domain knowledge is crucial for successful outcomes<br /></span></li> <li style="font-weight: 400;"><span style="font-weight: 400;">How to decide when your model is ready to go into  production and why you need to go beyond accuracy </span></li> </ul> <p><strong>Who is Taraneh Khazaei?</strong></p> <p><span style="font-weight: 400;">Taraneh Khazaei is</span> <span style="font-weight: 400;">Chief Data Scientist on the National AI Team,  Customer Success Unit at Microsoft. In this role, she advises Microsofts clients on how to adopt machine learning. Working with clients, Tara has</span> <span style= "font-weight: 400;">researched the state of the art of speech to text methods and technologies, developed deep sequential modeling methods (e.g., use of embeddings, RNNs, and transformer networks) on terabytes of clickstream data to model and predict user online behavior and designed and developed an ML pipeline to predict the market price of a vehicle. </span></p>
Takeaway 1 | 00:34 MIN
What it takes to be a good data scientist: 49% statistical intiution, 49% domain knowledge, 2% models
Takeaway 2 | 01:08 MIN
You can't throw data at an ML project, you need to take an iterative approach.
Takeaway 3 | 02:42 MIN
Bias and explainability are two of the biggest challenges AI projects face
Takeaway 4 | 01:59 MIN
Your data could reinforce societal biases.
Takeaway 5 | 02:01 MIN
How to make sure your model is ready for production.

“If you want to be a good data scientist, you should spend ~49% of your time developing your statistical intuition (i.e. how to ask good questions of the data), and ~49% of your time on domain knowledge (improving overall understanding of your field). Only ~2% on methods per se.” Nate Silver, a statistician and writer who analyzes sports, elections and more. 

In this week’s podcast Jon Prial is joined by Tara Khazaei, Chief Data Scientist, National AI Team, Customer Success Unit at Microsoft. Jon and Tara talk about how domain knowledge, as well as statistical intuition, make for more successful outcomes in machine learning projects. They discuss performance through the lens of projects Tara and her team have led at Microsoft. 

In this episode you’ll hear: 

  • Why you need enough, high-quality data 
  • The importance of iterating and validating your approach to achieve the best performance
  • The challenges bias and explainability pose to ML projects
  • Why domain knowledge is crucial for successful outcomes
  • How to decide when your model is ready to go into  production and why you need to go beyond accuracy 

Who is Taraneh Khazaei?

Taraneh Khazaei is Chief Data Scientist on the National AI Team,  Customer Success Unit at Microsoft. In this role, she advises Microsofts clients on how to adopt machine learning. Working with clients, Tara has researched the state of the art of speech to text methods and technologies, developed deep sequential modeling methods (e.g., use of embeddings, RNNs, and transformer networks) on terabytes of clickstream data to model and predict user online behavior and designed and developed an ML pipeline to predict the market price of a vehicle. 

More Episodes

Driving Digital Transformation using IoT and AI with Rob Rastovich

Data and Digital Ethics in ESG

ML Ops with Demetrios Brinkmann and David Aponte

The Story of AI Transformation at BMO

AI in Marketing with Paul Roetzer

AI Chatbots at Drift with Jeff Orkin