Tomasz Korulczyk, Ph.D.

The project code is the property of the company and for this reason, it is not publicly accessible.

Scripts utilizing the Prophet time series forecasting method for the extraction, preprocessing, and generation of forecast data. Applications include but are not limited to: sales predictions, forecasts of court hearings and judgments, and CHF/PLN exchange rate predictions.

Available at: https://github.com/tkorulczyk/codecool-personal-data-interceptor-ml

Open-source project developed as a part of the Codecool Upskill course. Scrpit aims to identify first and last names as personal data based on a deep learning model trained using TensorFlow. The model was trained on data from the dane.gov.pl database.

The project code is the property of the company and for this reason, it is not publicly accessible.

A service designed for supporting text preprocessing before passing it to OpenAI's GPT model. Its main goal is to identify and intercept personal data such as: phone numbers, emails, national identification numbers (PESEL), other sensitive data, bank account numbers, first names, surnames, postal codes, cities, and street addresses. The service offers several modes for handling this data:… Pokaż więcej

The project code is the property of the company and for this reason, it is not publicly accessible.

A service designed for supporting text preprocessing before passing it to OpenAI's GPT model. Its main goal is to identify and intercept personal data such as: phone numbers, emails, national identification numbers (PESEL), other sensitive data, bank account numbers, first names, surnames, postal codes, cities, and street addresses. The service offers several modes for handling this data: 1) Simple detection, 2) Simple replacement with a specified or default character sequence, 3) Substitution with randomly generated data, 4) Replacement with a random identifier that can be decoded later, 5) Encryption, specifically using the strongest encryption methods. Pokaż mniej

Available at: https://github.com/tkorulczyk/monday-crm-manager
For security reasons, the project is set to private. Please reach out to me, and I can grant access to recruiters by adding them as collaborators.

A hobby, made in free time, but comprehensive system for managing events and enhancing Monday.com's CRM for sales lead management. It boasts the following features:

Full communication: Complete CRUD operations for leads to and from Monday.com, sending and responding to… Pokaż więcej

Available at: https://github.com/tkorulczyk/monday-crm-manager
For security reasons, the project is set to private. Please reach out to me, and I can grant access to recruiters by adding them as collaborators.

A hobby, made in free time, but comprehensive system for managing events and enhancing Monday.com's CRM for sales lead management. It boasts the following features:

Full communication: Complete CRUD operations for leads to and from Monday.com, sending and responding to events executed in Monday.com using webhooks.
1) Automatic messaging: Automatic sending of SMS messages and formatted emails using Thymeleaf and the MimeMessage class, scheduled with cron - for instance, reminders for meetings or birthday SMS.
2) Personalized communication: Custom name declination for tailored messaging.
3) Appointment management: Registering and rescheduling sales meetings.
4) Inpost code dispatch: The ability to send Inpost codes to clients with a single click using the Inpost API.
5) Event-driven operations: Triggering and executing events with webhooks, such as sending business cards, sending meeting links, and dispatching emails with interview links for client feedback, and more.
6) Data integration: Fetching leads and updating statuses from another system via REST API (bi-directional communication). Pokaż mniej

The project code is the property of the company and for this reason, it is not publicly accessible.

A service designed for managing prompts appended to user queries for OpenAI's ChatGPT, aiming to specify a particular role and behavior of the model.

Available at: https://github.com/tkorulczyk/codecool-spaceship-titanic

As part of the Codecool Upskill course, I developed a script focused on a Data Science challenge related to the fictional "Spaceship Titanic" scenario. The core task was to utilize available data from the damaged spaceship's computer system to predict which passengers were affected by a spacetime anomaly.

To ensure the model's effectiveness, the process involved thorough data preprocessing (addressing missing… Pokaż więcej

Available at: https://github.com/tkorulczyk/codecool-spaceship-titanic

As part of the Codecool Upskill course, I developed a script focused on a Data Science challenge related to the fictional "Spaceship Titanic" scenario. The core task was to utilize available data from the damaged spaceship's computer system to predict which passengers were affected by a spacetime anomaly.

To ensure the model's effectiveness, the process involved thorough data preprocessing (addressing missing data, managing outliers, etc.), selecting an optimal classification model, and fine-tuning hyperparameters. Through these steps, I achieved an accuracy score of 0.80523. Pokaż mniej

The project code is the property of the company and for this reason, it is not publicly accessible.

Script that learns the characteristics of leads, such as: email length, email domain popularity, phone number length, digit arrangement, the presence or absence of a first name in the email or address details. Based on these characteristics, it classifies whether a lead is mistaken or genuine. Classification accuracy score is 94%

Avaliable at: https://github.com/tkorulczyk/codecool-recruitment-task

This script was developed as a qualifying recruitment task for Coolcode Upskill. The objective was to predict the spectroscopic signal values of sensors for isopropanol and water based on measured characteristics in air using a ML algorithm. Data was sourced from various files, consolidating 10 sensor measurements for air, water, and isopropanol. After ensuring data integrity, the dataset was segmented into test and… Pokaż więcej

Avaliable at: https://github.com/tkorulczyk/codecool-recruitment-task

This script was developed as a qualifying recruitment task for Coolcode Upskill. The objective was to predict the spectroscopic signal values of sensors for isopropanol and water based on measured characteristics in air using a ML algorithm. Data was sourced from various files, consolidating 10 sensor measurements for air, water, and isopropanol. After ensuring data integrity, the dataset was segmented into test and train subsets. Using linear regression, the independent variable (air) was analyzed against the dependent variables, achieving a model fit with 87.99% explained variance for training and 87.70% for testing. While air was a strong predictor, certain discrepancies were observed in the results, prompting a recommendation to explore non-linear regressions for potential improvements. Pokaż mniej

Available at: https://github.com/tkorulczyk/sda-weather-service

Java app developed as a part of the Java Fundamentals course offered by the Software Development Academy, this service is a straightforward tool that retrieves weather information from the Weather Stack API based on user inputs such as location and the desired date of weather data (current forecast or historical). It also provides the option to save the fetched weather data to an in-memory H2 database. The architecture of… Pokaż więcej

Available at: https://github.com/tkorulczyk/sda-weather-service

Java app developed as a part of the Java Fundamentals course offered by the Software Development Academy, this service is a straightforward tool that retrieves weather information from the Weather Stack API based on user inputs such as location and the desired date of weather data (current forecast or historical). It also provides the option to save the fetched weather data to an in-memory H2 database. The architecture of this solution is based on the clean architecture principle.

UPDATE: recently, the project has been updated with functionality that allows chatGPT to make recommendations on appropriate clothing and possible outdoor activities based on weather data. Pokaż mniej

Available at: https://github.com/tkorulczyk/sda-guess-number-game

Java app developed as part of the Java Fundamentals course offered by the Software Development Academy, this service is a basic game in which the user tries to guess a number randomly generated by the system. The game tracks the number of attempts and the time taken to guess the correct number.

Available at: https://github.com/tkorulczyk/sda-diary

Java app developed during the Java Fundamentals course offered by the Software Development Academy. App serves as a basic electronic diary. It allows users to save entries with a title and content to a database and provides functionalities for both writing and retrieving these entries.