Pavel Logacev
Verified Expert in Engineering
Data Scientist and AI Developer
Berlin, Germany
Toptal member since January 27, 2022
Pavel is a data scientist specializing in Bayesian methods. He has a master's degree in computational linguistics and a PhD from Potsdam University in Germany. Pavel has over 10 years of experience in statistical data analysis and data science, having worked in sectors as diverse as pricing, psychology, finance, education, health, eCommerce, SEO, and betting markets.
Portfolio
Experience
- R - 15 years
- Ggplot2 - 10 years
- Data Science - 7 years
- Bayesian Inference & Modeling - 7 years
- Machine Learning - 5 years
- Statistics - 5 years
- Stan - 5 years
- Python - 5 years
Work Experience
Principal Data Scientist - Optimization
Centric Software
- Furthered the development of the price optimization engine.
- Furthered the development of demand and price-response models.
- Quantified the value of the price optimization engine.
Founder
Sell Smart
- Developed an automatic price optimization model based on sales data with regular stockouts.
- Developed a demand forecast model based on sales data with regular stockouts.
- Built an inventory alerting mechanism based on a probabilistic demand forecast model.
- Implemented an autonomous pipeline that automatically alerted Shopify merchants if they risked running low on stock.
Data Scientist
Freelance
- Implemented machine learning solutions that helped clients make sense of data and automate forecasting processes.
- Provided statistical consulting and carried out statistical analyses, contributing to scientific publications.
- Solved numerous continuous and discrete optimization problems.
Senior Data Scientist
Pearson Pricing
- Carried out price elasticity analyses on big data to determine the optimal price points for stores and products and presented the results.
- Conducted basket analyses on big data to determine popular combinations of items customers tend to purchase.
- Created a standardized workflow for price elasticity analysis in BigQuery, which can be applied to large datasets from various industries.
- Trained analysts in the use of R for data analysis.
Assistant Professor
Bogazici University
- Designed and taught courses on research methods and statistics.
- Created new statistical models of eye movements during sentence comprehension in reading.
- Planned and carried out experimental research on reading and managed an eye-tracking laboratory.
- Supervised the research of several students, including master's theses.
Data Scientist and R Developer
Procter & Gamble
- Designed and implemented a custom algorithm to automatically determine the main drivers of change in sales and automatically report the key changes.
- Developed the computational architecture and infrastructure for the analytics solution.
- Managed a team of 4 data scientists in developing the product.
Researcher
University of Potsdam
- Taught courses on experimentation and research methods.
- Conducted research on formal models of sentence comprehension.
- Managed an SQL database with linguistically annotated historical texts.
Software Developer
adisoft AG
- Developed mobile communication solutions on embedded platforms like Symbian operating system.
- Built web applications in Python for the Zope CMS.
- Restructured and ported software from Linux to Symbian operating system.
Project History
Energy Consumption Prediction
Based on weather data, I created a REST API for forecasting the daily by-block energy demand distribution. The quantile regression model was implemented on H2O, and the REST API in R, using the plumber library.
The predicted distribution of energy consumption was used in simulations serving to optimize the energy efficiency of the district heating system.
Genetic Programming-based Trading System
I developed and implemented the logic of the optimization algorithm in R and backtested on historical data. Signals were exported in a MetaTrader readable format.
Churn Prediction App
The app displays predictions of churn probability and features importance values.
Price Optimization for a Major Fast-food Chain in Europe
I built a multi-level elasticity model using POS data, capturing nonlinear price-response relationships in log-log space, and explicitly estimating elasticities and cross-elasticities between different meal sizes. This enabled us to jointly optimize prices for each size triplet, balancing upsell incentives and margin targets. After scenario analysis and alignment with the client’s pricing team, we delivered an actionable optimized price list—about 70% of which was adopted chain-wide. The entire project, from model design to stakeholder handover, was completed in six weeks.
Sell Smart: Price Optimization
http://sell-smart.appEducation
PhD in Cognitive Science
University of Potsdam - Potsdam, Germany
Master's Degree in Computational Linguistics
University of Potsdam - Potsdam, Germany
Skills
Libraries/APIs
Ggplot2, Pandas, NumPy, Tidyverse, Scikit-learn, Joblib, PyMC, Rcpp, Matplotlib, XGBoost, PyTorch, React, TensorFlow, Shopify API, OpenCV, SpaCy
Tools
ARIMA, SARIMA, StatsModels, ChatGPT, Claude Code, Git, BigQuery, Prefect, Pytest
Languages
R, Python, SQL, C++, Perl, C, JavaScript
Industry Expertise
Applied Statistics, Retail & Wholesale
Frameworks
LightGBM, RStudio Shiny, Streamlit
Paradigms
ETL, B2B, REST
Platforms
Google Cloud Platform (GCP), Ubuntu, H20, RStudio, Amazon Web Services (AWS), Docker, Vertex AI
Storage
Databases, MySQL, Data Pipelines, PostgreSQL, Google Cloud Storage, JSON
Other
Linguistics, Statistics, Machine Learning, Cognitive Science, Data Science, Scientific Data Analysis, Data Visualization, Data Analysis, Hypothesis Testing, Regression, Data Analytics, Statistical Data Analysis, Statistical Modeling, Analytics, Regression Modeling, Linear Regression, Statistical Analysis, Modeling, Statistical Methods, Full-stack Development, Logistic Regression, Pricing Models, Data, Supervised Learning, Feature Engineering, Pricing Elasticity, Data Preprocessing, Model Development, Random Forests, Natural Language Processing (NLP), Bayesian Statistics, Bayesian Inference & Modeling, Stan, A/B Testing, Time Series, Predictive Modeling, Software Development, Software Engineering, Mathematics, Predictive Analytics, Forecasting, Data Engineering, Big Data, Sales Forecasting, Pricing, Cloud, Fantasy Sports, Applied Mathematics, Data Management, Presentations, Product Forecasts, Demand Forecasting, Time Series Forecasting, Time Series Analysis, Unsupervised Learning, Dynamic Pricing, Gradient Boosting, ETL Pipelines, Model Validation, Probabilistic Modeling, OpenAI, Pricing Optimization, Pricing Strategy, Large Data Sets, Data Cleaning, Datasets, Communication, Model Deployment, Model Evaluation, Causal Inference, Computational Linguistics, Genetic Algorithms, Google BigQuery, dbplyr, Business Analysis, Sports, Big Data Architecture, Cost Reduction & Optimization (Cost-down), Revenue Optimization, Generative Pre-trained Transformers (GPT), NumPyro, Artificial Intelligence (AI), eCommerce, API Integration, Live Pricing, Spatial Analysis, Polars, Computer Vision, Data Labeling, Large Language Models (LLMs), Object Detection, Stakeholder Management, Deep Learning
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring