Pie & AI: Kyiv - From Data to Prediction

Event Overview

Pie & AI Kyiv Meetup: From Data to Prediction is a practical introduction to how machine learning models learn from data and generate predictions.
The session will introduce the core concepts behind neural networks and demonstrate them through a hands-on financial market data example using Python and TensorFlow/Keras.
Participants will follow the machine learning workflow from data preparation and feature selection to model training, prediction, visualization, and evaluation. A short additional demonstration will show how predictive analytics can also be approached using SQL and Google Cloud BigQuery ML.
Financial market data will be used as an educational case study to demonstrate machine learning concepts rather than as a method for providing investment advice or claiming reliable prediction of future market prices.
The meetup will conclude with an open discussion and Q&A about model performance, data quality, limitations, and practical applications of machine learning.

Pie & AI is a series of DeepLearning.AI meetups independently hosted by community groups. This event is hosted by Valentyn Verovkin. Special thanks to their support!

Event Agenda & Speakers

16:00–16:05 — Welcome and Pie & AI Introduction
Introduction to the meetup, its objectives, and today's learning journey.
16:05–16:15 — From Traditional Programming to Machine Learning
How machine learning changes the conventional relationship between data, rules, and outcomes.
16:15–16:25 — Understanding Neural Networks
A beginner-friendly introduction to how neural networks learn patterns from data.
16:25–16:35 — From Financial Data to an ML Problem
Preparing financial market data and defining inputs, targets, training data, and test data.
16:35–16:50 — Practical Demo: TensorFlow/Keras
Building, training, and evaluating a simple neural network using financial market data.
16:50–17:00 — Machine Learning with SQL: BigQuery ML
A short demonstration of an alternative SQL-based approach to predictive analytics using Google Cloud.
17:00–17:10 — What Does a Prediction Actually Tell Us?
Model evaluation, limitations, data quality, overfitting, and responsible interpretation of predictions.
17:10–17:15 — Open Discussion and Q&A

Valentyn Verovkin — Organizer and Speaker
From Data to Prediction: Practical Machine Learning
The session will cover the transition from traditional programming to data-driven machine learning, basic neural network concepts, and a practical TensorFlow/Keras demonstration using financial market data. A short comparison with SQL-based predictive analytics using Google Cloud BigQuery ML will illustrate an alternative approach to working with machine learning models.
The speaker's academic and professional interests include data analytics, machine learning, and financial engineering.
LinkedIn: https://www.linkedin.com/in/valentyn-verovkin-79606934/

________

For learners of all levels; Beginner

Language: Ukrainian

About Pie & AI: Pie & AI is a series of DeepLearning.AI meetups independently hosted by our global AI community. Events typically include conversations with leaders in the world, thought-provoking discussions, networking opportunities with your fellow learners, hands-on project practice, and pies (or other desserts you prefer.)