Infinite Curiosity Pod with Prateek Joshi - AMA with Prateek Joshi: Space industry, synthetic data in machine learning, foundation models, AI-infused coding tools

AMA with Prateek Joshi: Space industry, synthetic data in machine learning, foundation models, AI-infused coding tools

06/02/22 • 14 min

Infinite Curiosity Pod with Prateek Joshi

In this AMA episode, the host Prateek Joshi answers the following questions:
- How is machine learning being used in the space industry?
- How is synthetic data being used to train AI systems?
- Are foundation models going to become more prevalent in production ML systems?
- What do you think about AI-infused coding tools?

Previous Episode

undefined - Serg Masis on interpretable machine learning, process fairness vs statistical fairness, how to measure interpretability, how to interpret neural networks, how to increase the interpretability of a model

Serg Masis on interpretable machine learning, process fairness vs statistical fairness, how to measure interpretability, how to interpret neural networks, how to increase the interpretability of a model

Serg Masis is a Data Scientist in agriculture with a background in entrepreneurship and web/app development. He's the author of the book "Interpretable Machine Learning with Python". In addition to ML interpretability, he's passionate about explainable AI, behavioral economics, and ethical AI.
In this episode, we cover a range of topics including:
- How did he get into machine learning?
- What is interpretable ML?
- What is post hoc interpretability?
- Process fairness vs statistical fairness
- How an algorithm creates the model?
- How a model makes predictions?
- What makes an ML model interpretable?
- How do you measure the interpretability of a model?
- How do parts of the model affect predictions?
- Does the method of interpretation depend on the model? Or can we apply a given method to a number of models?
- Can you explain a specific prediction from a model?
- What techniques can we use to interpret neural networks?
- What techniques are available to increase the interpretability of a model?

Next Episode

Vin Vashishta on injecting reality into machine learning products and creating business value with data science

Vin Vashishta is a globally recognized expert on AI Strategy and Data Science. He has been a LinkedIn Top Voice in Data Science and has been featured on dozens of Top 10 Lists over the last 7 years. His client list includes the likes of Walmart, JPMC, Siemens, and Airbus. He delivers products with recurring revenue streams in the 100s of millions and build Data Science teams from the ground up. He advises startups and teaches founders how to launch their first ML based product. He founded V-Squared in 2012 and built it into a successful AI Strategy consulting practice. 4 years ago, he started teaching Business Strategy For Data Scientists.
In this episode, we cover a range of topics including:
- Teaching a strategy class for data scientists
- How to measure the success of data science projects
- How to identify the highest value opportunities
- You learn a lot by building ML models, but you learn more my maintaining them for 6 months. Why is that?
- What are transferable capabilities for a professional who wants to transition into data science?
- How to inject reality into ML products using domain knowledge?
- Advantages of rapid prototyping
- How to structure AI strategy sessions with potential new collaborators?
- What is career coaching?
- How to hire great data scientists?