Colin Priest

#chatgpt can assess whether you are a good match for an advertised job and recommend how to improve your ATS score. 1) Load up the job advertisement and your CV to ChatGPT 2) Use this prompt: #resumeoptimization #chatgpt #atsfriendly

Understanding and adapting to customer behavioral changes is a cornerstone of modern business strategy. In this 2-minute video, I share a cool #featureengineering trick to help quantify how much a customer’s behavior has changed. #customersegmentation bit.ly/3qAQkm6

Want to know who your most valuable customers are? Research shows that clumpiness signals predict who your most valuable customers will be. featurebyte.com/resources/feat… Try the Python code for yourself at @DataCamp app.datacamp.com/workspace/w/b1… #customersegmentation #featureengineering #ai

Feature engineering is a crucial step in ML; it's how you improve the predictive performance of your models. @ColinAPriest, Chief Evangelist at FeatureByte, covers managing the features for your ML models, saving time and improving consistency. 🔗👇 ow.ly/TzaU50PrEhn

Almost every day, I see a report of how someone found a way to hack an AI system or an AI was using the wrong data. #mlops #ai featurebyte.com/resources/mlop…

Fraudsters follow a playbook, a well-defined process of particular amounts and transaction descriptions. In this 2-min video, I share feature engineering ideas to help detect fraud using similarity signals. featurebyte.com/resources/feat… #featureengineering #frauddetection #datascience

The rate of change in #ai this year is off the scale! How do people keep up with what's available? How do people manage version control? Enterprises worry that their #genai and #featureengineering pipelines have no version control. featurebyte.com/resources/mlop…

If you're a bank or fintech modeling #creditrisk, it turns out that there are a few tricks to #featureengineering to capture change signals. Learn how in this 2-minute video featurebyte.com/resources/feat… #datascience #ai

The regularity of a customer's behaviors is an indicator of their risk aversion and money management. Watch this two-minute video to see how to improve your credit risk models using entropy as a regularity metric. #customersegmentation #featureengineering bit.ly/44OeoQB

We all know that customer loyalty is good for business. But what about the breadth and depth of that loyalty? What about product loyalty versus brand loyalty? Watch my video explaining how to feature engineer for loyalty using entropy. #featureengineering bit.ly/3D9k2B4

My wife talking me into seeing a musical while we visit New York, but I'm not into the Disney musicals she likes. Then she mentioned Back to the Future - The Musical! And that got me thinking about time travel and data science. featurebyte.com/resources/mlop…

#pandas is great for #featureengineering but it doesn't scale. #SQL is great for scale but time-aware SQL is difficult to maintain. The solution: automatically generated SQL and Spark for #ai ready data featurebyte.com/resources/mlop…

Today Singapore celebrated Hari Raya Haji, the Muslim pilgrimage to Mecca known as the haj. Even for the rest of us, travel is a great way to expand our understanding of life. But the same doesn't apply to data. Moving data is expensive and dangerous. featurebyte.com/resources/mlop…

If you're in Singapore on 14th July, I will be at the Singapore Actuarial Society's afternoon forum, talking about my journey from actuary to data scientist. You can register at actuaries.org.sg/events/afterno… #actuary #datascientist #careerchange

Individuality is important to Aussies like me. Yet marketing puts me into a group with millions of others and treats us like we are all the same. In this short video, I show how to feature engineer for dissimilarity signals. #featureengineering bit.ly/44p1Qzb

If you work around pandas' in-memory aggregation RAM limitations, you run into speed limitations. It's 700 times slower! Maybe it's time to push aggregation calculations into the database... #featureengineering #pandas #ai featurebyte.com/resources/feat…

Neerja Bhanot was a flight attendant on Pan Am Flight 73 when it was hijacked in Pakistan by Palestinian terrorists on September 5, 1986. She helped the pilots escape, saved American passengers from execution, and opened the emergency door so hundreds could jump to safety. Ultimately, she was killed while shielding three children from gunfire. Twenty-two-year-old Indian air hostess Neerja Bhanot was always a woman of courage and convictions. Despite the traditions of her culture and her family's protests, she left her arranged marriage when her husband proved abusive. Then she struck out on her own and began working as both a model and a flight attendant for Pan Am. But shortly after she began her new career in the skies, disaster struck when a flight she was working was hijacked by four terrorists in 1986. Though she was new to the job, she proved to be the greatest hero of the entire ordeal. Over the course of 17 harrowing hours, she helped both the pilots and hundreds of passengers to escape. In the end, she was killed while using her body to protect three children from gunfire. Bhanot was posthumously awarded medals of bravery from India, Pakistan, and the United States. And one of the children she saved grew up to be a pilot himself and credited Bhanot for his decision.

Want to improve your fraud detection models? Feature engineer for stability signals. In this quick video, I show how. #frauddetection #ai #featureengineering featurebyte.com/resources/feat…

Feeling the pain of increasing cloud costs? Here's how organizations can reduce AI infrastructure costs without compromising the success of their projects. #ai #cloudcostoptimization aijourn.com/are-ai-infrast…

#pandas aggregations don't scale because they are too RAM intensive. In this simple example, I show how a 200k dataset requires 6GB of RAM for a windowed aggregation! #ai #featureengineering featurebyte.com/resources/feat…