Flexport: The Pragmatist’s Guide to Applying Machine Learning | TransformX 2022

Want to take advantage of the latest research and advances in ML? Looking to chase the next big thing? Often the most useful machine learning models are based on the pain points you see every day. Tom Vu, Senior Director and Head of Data Science and Machine Learning at Flexport, has identified and applied disruptive machine learning opportunities that resulted in over $2 billion of value during the course of his career.
In this keynote, Vu will share his insights on how the limitations posed on human reasoning help identify ML opportunities, the importance of understanding processes and pain points relative to cognitive load, and what led him to decide to build an AI model to predict costs from CAD drawings. Vu’s portfolio of applied research projects and interests include routing, scheduling, assignment optimization under uncertainty, geospatial-temporal forecasting, imperfect information games, natural language processing, and computer vision. Prior to Flexport, Vu was the Head of Data Science and Analytics at WeWork, and the Chief Data Scientist at Boeing. He has over 20 years of experience implementing vision, transforming unmet business opportunities into realized software solutions.
👉 Check out more here: scl.ai/3z0SzQG

[@Dannesito]

Hola soy de Ecuador, me inspira mucho todas estas industrias y grandes emprendimientos de tecnología, crean soluciones fantasticas, gracias por crear cambios para nuestro mundo y nuestra vida, Dios los bendiga.

[@happybydefault]

Thank you, Tom and people at Scale AI.
The following is a summary of this video, from OpenAI's Whisper and GPT-3:
“The speaker, Tom Boo, is a machine learning scientist and software innovator. He is currently the senior director and head of data science and machine learning at Flexport. Tom's portfolio of applied research projects and interests include routing, scheduling, and assignment optimization under uncertainty, geospatial temporal forecasting, imperfect information games, natural language processing, and computer vision. Tom has over 20 years of experience implementing vision, transforming unmet business opportunities into realized software solutions.
“Tom starts by talking about how in an information rich world, it's easy to drown in information while thirsting for knowledge. He then shares a story from his career of how he was able to deliver value in a previous job by using machine learning to help a company improve their cost efficiency of manufacturing. Tom and his team created algorithms that could take the design of a complicated piece of machinery and predict how much it would cost to manufacture that, giving procuring agents the ammunition they needed to negotiate favorable prices for parts, and as a result, bring the production costs down.
“Tom talks about how the value of machine learning lies in the opportunities that suffer from bounded rationality, which is the limitations imposed on human reasoning that's based off of how complicated the problem is, how tractable the problem is, combined with how much time you have to solve that problem. He gives the example of Frederick Winslow Taylor, who was an industrial laborer and machinist who was one of the first people to apply the principles of data collection, scientific analysis, and optimization to the work of human labor of industrial production. Tom talks about how a lot of the tools and analysis from the past cannot be applied in the same way today because of the near exponential increase in the amount of data stored on the internet.
“In conclusion, Tom talks about how he believes that the best opportunities for machine learning are in the areas where cognitive load is taxing for people, and that the goal should be to create solutions that will simplify and make human reasoning more effective and efficient.”

[@rajganesh9095]

Freelance remote job India

[@olegkalinkin6877]

Next time please try some machine learning tools to clean up the audio. Or better yet get decent recording equipment.