#73 Powering Change using AI with Alex Ermolaev – AI Leader

Alex Ermolaev.jpg

Alex Ermolaev has been involved in the software industry for 20 years, including AI-specific experience at Bell Labs, Microsoft, several startups and now Nvidia. He is currently a leading AI software developer and works with groundbreaking companies that are implementing incredible AI solutions across several domains. 

In this episode, Alex describes how he started in the data space. Early in his career, he got a chance to work on a lot of data and software products. His first application that he enjoyed was expert systems. Expert systems are an interesting technology; however, it worked quite well in very few scenarios. Sometimes people spent a lot of time and budget on these systems with no return. Please don’t assume that technology will be able to do everything you need it to do before spending time and money. Alex came to the United States for an internship at Bell Labs, then worked for Microsoft, and startups where he did several AI projects. Some people try to use AI to ensure a TV advertisement is shown at the appropriate time or not, whereas some people will use it for brand logo stuff. About five years ago, AI started getting much attention again. It is possible to achieve higher accuracy with AI systems than anyone was able to see before. Because of this, Alex decided to dedicate all of his time to AI projects and nothing else.  

Then, Alex and Felipe discuss business development skills. It is easy to sit in an office and assume the world works in a certain way. The only way to understand how the world truly works is by getting out and experiencing it. He went for a business trip and saw so many new things, got new information, and made new contacts. When he comes back to the office, nothing has changed. You realize the company will operate in a certain way despite all the changing aspects outside of those walls. Sometimes the company will see business development as an external process. When Alex was looking for a company who was taking advantage of the AI resurgence, he came across Nvidia. Thousands of people are working with AI for Nvidia now. Alex helps organizations across the board adopt the tools of AI for their particular applications. He loved working and interacting with people trying to solve problems. 

Then, Alex speaks about the successes and failures he saw when working for companies and their AI technology. Alex says AI is finding patterns in data to predict or detect something. There are many cases where patterns do not exist. For instance, in stock markets, patterns tend to be very weak because changes in the stock market tend to be very external. People try to apply AI to the stock market, but the pattern is very weak. Computer viruses and fraud are examples of successful AI patterns. Fraud tends to happen in patterns, whether it is telecom fraud or e-commerce fraud. If you detect a pattern, you can apply it to future cases to stop fraudulent activities. Stay tuned to hear Alex discuss how to tackle data problems using AI, some exciting uses for AI, and what Alex is most proud of in his career. 

Enjoy the show!

We speak about:

  • [01:50] How Alex started in the data space

  • [04:55] Alex’s professional background  

  • [10:30] Working for the finance team at Microsoft 

  • [14:55] Business development skills  

  • [18:50] Challenges working with startups 

  • [22:10] Working at Nvidia  

  • [26:20] Successful and unsuccessful AI patterns  

  • [30:00] AI and collecting data 

  • [35:15] How to tackle data problems using AI 

  • [40:30] Exciting uses for AI  

  • [43:15] The execution of new AI programs  

  • [49:00] What Alex is most proud of  

  • [50:20] Be patient and invest in your knowledge   

Resources:

Alex’s LinkedIn: https://www.linkedin.com/in/alexermolaev


Quotes:

  • “The best way to develop knowledge in any area is to experience it.”

  • “It is easier to sit in an office and assume the world works in a certain way.”

  • “Don’t be in startups because it’s cool, try and find a path that meets your own needs.”

  • “Working with startups is a lot of broader outreach and helping the community understand what is possible.”

Thank you to our sponsors:

UNSW Master of Data Science Online: studyonline.unsw.edu.au

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au

Fyrebox - Make Your Own Quiz!

Alex Ermolaev is based in San Francisco, California.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#72 Focusing on Simplification to Solve Data Science Roadblocks with Evan Shellshear – Head of Analytics

Evan Shellshear.jpg

Evan Shellshear has been an entrepreneur for more than a decade, and throughout that time he has always loved getting his hands dirty with building products from scratch and then commercializing them. Evan has a passion for innovation and not just from a managerial perspective but also from a doing perspective. He has a Ph.D. in Game Theory, is published in fields computer graphics to politics, mathematics to manufacturing, and much more. Evan has founded or co-founded over half a dozen companies to commercialize different technologies. 

In this episode, Evan explained how he started in the data space, including his Ph.D. in Game Theory and hitting rock bottom in Germany. Then, Evan describes what to do if you are stuck in the details and you have hit a roadblock, and you cannot go any further. Imagine running around on a piece of paper, and someone draws a line in front of you, the best solution you can do is to zoom out and see how you are in a three-dimensional world. The best way to solve technical roadblocks is by zooming out and asking what the problem is you are trying to solve. If you still cannot solve the barrier, zoom out another level, and ask what the client is trying to achieve. Instead of solving by being technically smarter, you can solve it by understanding the problem better. 

In Sweden, Evan worked with robotics and took a different approach than the rest of the world to solving deep fundamental issues in robotics around path planning. Instead of using conventional techniques, they chose to open up and find all new avenues of exploration. One thing data scientists are taught is to be highly technical; they are not taught how to learn things. Whenever Evan lectures, he tells people, however complicated their first solution is, look at it and simplify it, then simplify it again. It needs to be simplified twice before then can present it to anybody. Then, Evan explains the measures of success he uses during a project to ensure he is on the right track. When you notice your client has not even considered the problem you are presenting is what Evan calls an “aha moment.” Later, Evan explains the process of a case study, getting users to adopt new technologies and his book Innovation Tools.

Enjoy the show!

We speak about:

  • [01:15] How Evan started in the world of data

  • [09:45] Zoom out to solve technical roadblocks 

  • [12:10] Examples of how Evan zoomed out

  • [14:25] Why is zooming out a challenge for data scientists?   

  • [18:00] Focus on simplification

  • [22:45] Taking opportunities that present themselves   

  • [27:00] Measures of success during a project 

  • [31:00] The process of a case study 

  • [36:05] Getting users to adopt new technologies  

  • [40:00] Innovation Tools

  • [47:20] Evan’s proudest moment 

  • [49:40] Challenges for the future of machine learning  

  • [51:30] Get soft skills   

Resources:

Evan’s LinkedIn: https://www.linkedin.com/in/eshellshear/

Innovation Tools: https://amzn.to/2OrrAsj

Quotes:

  • “Take a step up and over to look at the problem in a new direction.”

  • “It is in our human nature to overcomplicate things.”

  • “I need to help the company understand what the true problem is.”

  • “Take a low-risk approach to solve your client’s problem.”


Thank you to our sponsors:

UNSW Master of Data Science Online: studyonline.unsw.edu.au

Datasource Services: datasourceservices.com.au or email Will Howard on will@datasourceservices.com.au

Fyrebox - Make Your Own Quiz!

Evan Shellshear is based in Kensington, Victoria, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!

#70 Making Black Box Models Explainable with Christoph Molnar – Interpretable Machine Learning Researcher

Christoph Molnar.jpg

Christoph Molnar is a data scientist and Ph.D. candidate in interpretable machine learning. He is interested in making the decisions from algorithms more understandable for humans. Christoph is passionate about using statistics and machine learning on data to make humans and machines smarter.

In this episode, Christoph explains how he decided to study statistics at university, which eventually led him to his passion for machine learning and data. Starting out studying with a senior researcher gave Christoph exposure to many different projects. It is an excellent program for students and companies whom both benefit greatly. Christoph learned so much about statistics that he would not have been able to acquire otherwise. The clients got nine hours of consulting for free, which is very valuable for their businesses. When Christoph started his statistical consulting career, he did patient analysis to assess if a medication was affecting the spine. He found this very interesting as it differed significantly from his previous consulting.

When labeling data, Christoph says to label and always compare continuously. For instance, when a student labeled one photo, later on, Christoph would show a student the same photo and see if it got labeled identically. Sometimes people will see the same image but label it differently; so, this is one thing you can do to ensure labeling data is going smoothly. If you have multiple labelers, you will need to compare how each labeler will mark the same photo. Do not be blind to the quality of your data; it is easy to adjust the numbers. 


Then, Christoph speaks about pursuing his Ph.D. in Interpretable Machine Learning. He publishes his book, Interpretable Machine Learning, on his website chapter by chapter. Christoph gets feedback and uses it while continuing his writing on future chapters. Learning about interpretable machine learning is not exactly present at university now. Some schools and professors are starting to integrate it into the curriculum. Stay tuned to hear Christoph discuss accumulated local effects, deep learning, and his book, Interpretable Machine Learning.

Enjoy the show!

We speak about:

  • [02:10] How Christoph started in the data space

  • [09:25] Understanding what a researcher needs

  • [15:15] Skills learned from software engineers 

  • [16:00] Statistical consulting 

  • [19:50] Labeling data  

  • [23:00] Christoph is pursuing his Ph.D.

  • [29:00] Why is interpretable machine learning needed now? 

  • [31:00] Learning interpretability  

  • [33:50] Accumulated local effects (ALE)

  • [37:00] Example-based explanations  

  • [39:15] Deep learning  

  • [43:35] The illustrations in Interpretable Machine Learning.

  • [49:50] How Christoph maximizes the impact of his time

Resources:

Christoph’s LinkedIn: https://www.linkedin.com/in/christoph-molnar-63777189/
Christoph’s Website: https://christophm.github.io

Interpretable Machine Learning: https://christophm.github.io/interpretable-ml-book/

Quotes:

  • “Always look at the process when labeling data.”

  • “After each chapter of my book, I publish it and get feedback.”

  • “I randomly read a lot of papers and structure the knowledge to fit them together.”

  • “I express what I want easier with illustrations in my book.”


Christoph Molnar is based in Munich, Bavaria, Germany.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!


#67 12 Questions to Ask in Mentoring Sessions with Felipe Flores – Founder & Podcast Host

Felipe Flores.jpg

In this episode, Felipe talks about an aspect of leadership, the one-on-one mentorship and feedback session with people on your team. To start, ask your team member what is on their mind and in general, how have things been going? If they do not have anything pressing they want to discuss in the sessions, Felipe turns to set of twelve questions. The first question is what you are most proud of that you have done since we last caught up? These questions are designed as tools for the team member to recognize the need for self-assessment. What they say is not essential; really, the follow-up questions are more necessary to help them uncover themselves. Next, ask what the team member could have done better since the last time you talked. By allowing them to evaluate and think of improving continually, they can learn faster and become more efficient in their work. 

The next few questions require a broader perspective and ask about the team as a whole. It should be clear that everything is a team effort, and all member’s ideas are respected and heard by others on the team. After the team questions, head back to questions about the person and ask what they would like to work on or improve? Then, the next issue will take a lot of trust and rapport with your team member, ask what is one thing that is true that you think I do not want to hear? Question nine is how I can help you to do better? This question has taught Felipe that he is good at the big picture but needs to focus on the details and how the team might achieve the big picture. Finally, the last three questions are asking what they like best and least about the organization and if they are happy at the moment.

Enjoy the show!

We speak about:

  • [01:35] Open-ended questions

  • [02:15] What are you most proud of that you have done since we last caught up?

  • [03:45] What could you have done better?  

  • [05:30] Questions about the team

  • [10:00] What would you like to work on or improve? 

  • [11:40] What is one thing that is true that you think I do not want to hear?

  • [13:10] How can I help you to do better?

  • [15:00] What do you like the least about the team? 

  • [16:00] What do you like best about the team?  

  • [16:35] Are you happy at the moment? 

Resources:

Saturday Night Live

Quotes:

  • “One of the best and quickest ways to learn is to evaluate your efforts continually.” 

  • “Understand the human behind the data scientist.” 

  • “You don’t need to be fixing every person’s problem, but everyone needs help every now and again.”

Thank you to our sponsors:

Fyrebox - Make Your Own Quiz!

RMIT Online Master of Data Science Strategy and Leadership

Gain the advanced strategic, leadership and data science capabilities required to influence executive leadership teams and deliver organisation-wide solutions.

Visit online.rmit.edu.au for more information

Felipe Flores is based in Melbourne, Australia.

And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. Thank you so much for listening. Enjoy the show!