Skip to content
View AndreyLoskutov's full-sized avatar

Block or report AndreyLoskutov

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AndreyLoskutov/README.md

Hi, I'm Andrey Loskutov

Senior Product & Marketing Analyst with 12+ years turning data into revenue across gaming, fintech, HR-tech, and staffing.
Expanding senior analytics experience into applied ML, experimentation, analytics engineering, and GenAI.

Based in Charlotte, NC | Green Card | Open to U.S. opportunities

LinkedIn Website Hugging Face Email LeetCode


What I Do

I build SQL-first product and marketing analytics: funnels, cohorts, LTV/ROAS, A/B testing, executive KPI systems, and stakeholder-ready recommendations. My background includes analytics leadership at G5 Entertainment and Sberbank HR-tech, plus current U.S. staffing analytics work.

Over the last two years, I have expanded into applied ML and GenAI where they improve decision quality: calibrated churn probabilities, cost-aware thresholds, source-grounded RAG, explainability, testing, and deployable Python apps.

Open to roles: Senior Data Analyst | Senior Product Analyst | Data Scientist | Analytics Engineer | Applied AI / GenAI Analytics roles


Featured Data Science Portfolio

Project What it shows Verified result Demo Repo
Telecom Customer Churn Prediction Calibrated churn scoring, thresholding, SHAP, lift analysis, Streamlit deployment Test ROC-AUC 0.8397; PR-AUC 0.6551; recall 0.9278 at threshold 0.12; top-decile lift 2.828 Demo Repo
A/B Test Analyzer Product experimentation, hypothesis testing, Bayesian cross-checks, rollout guardrails Conversion increased from 42.0% to 66.0% (+24.0 pp, p=0.008026); time-on-page increased by 1.691 minutes (p=0.000139) Demo Repo
Medical Q&A RAG Assistant Source-grounded RAG, citations, safety fallback, inference-only GenAI deployment RAGAS across 20 questions: answer relevancy 0.9644; context precision 0.9278; context recall 0.6067; faithfulness 0.5735 Demo Repo
SuperKart Revenue Prediction Retail revenue regression, prediction intervals, SHAP, Streamlit/FastAPI deployment RMSE $281.04; MAPE 4.23%; 73.7% RMSE reduction vs naive baseline; 85.7% interval coverage Demo Repo
Hotel Booking Cancellation Risk Scoring Hospitality risk scoring, cost-aware thresholding, SHAP, revenue-protection framing Test ROC-AUC 0.9172; PR-AUC 0.8385; recall 0.9167 at threshold 0.31; top 30% by risk captures 77% of cancellations Demo Repo

Recommended recruiter order: Telecom Churn, A/B Test Analyzer, Medical RAG first. SuperKart and Hotel are strong secondary projects with live demos.


Additional SQL Analytics Work

These projects support the analytics-engineering and senior analyst side of my profile:

Project area Signal Repo
A/B testing in SQL Experiment design, metric definition, statistical thinking Repo
Retention cohorts Cohort logic, repeatable retention views, product analytics Repo
LTV / ROAS funnels Growth analytics, marketing efficiency, acquisition quality Repo
Feature engineering for ML SQL-to-ML data preparation and reusable feature logic Repo
ETL / Redshift-style pipeline Data modeling, warehouse thinking, reproducibility Repo

Tech Stack

Analytics and BI

SQL Python pandas NumPy Tableau Power BI Looker Studio

Statistics and Experimentation

scipy statsmodels A/B Testing Power Analysis

Databases and Warehouses

PostgreSQL ClickHouse MySQL Redshift

Machine Learning

scikit-learn XGBoost LightGBM Optuna SHAP

GenAI and LLM

LangChain ChromaDB OpenAI Hugging Face RAGAS

Deployment and MLOps

Streamlit Gradio FastAPI Docker GitHub Actions AWS


How I Think About Analytics

  • The decision layer matters more than the model score. Calibrated probabilities, cost-aware thresholds, and a clear action path turn analysis into leverage.
  • Experiment results need guardrails. A statistically significant conversion lift still needs retention, revenue, performance, and novelty-effect checks before broad rollout.
  • Cohorts beat snapshots. Retention curves, LTV by acquisition month, and funnel-stage cohorts reveal stories that aggregate KPIs hide.
  • Production hygiene matters. If a model is meant to influence decisions, it needs reproducible code, saved artifacts, tests, documented limitations, and a usable interface.

Get In Touch


For recruiters: project map by role
Target role Best projects to review first
Senior Data Analyst A/B Test Analyzer; Telecom Customer Churn; Hotel Booking Cancellation
Data Scientist Telecom Customer Churn; SuperKart Revenue Prediction; Hotel Booking Cancellation
Product Analyst A/B Test Analyzer; Telecom Customer Churn; Hotel Booking Cancellation
Analytics Engineer SuperKart Revenue Prediction; Telecom Customer Churn; SQL analytics projects
ML Engineer Telecom Customer Churn; SuperKart Revenue Prediction; Medical Q&A RAG Assistant
Applied AI Engineer Medical Q&A RAG Assistant; Telecom Customer Churn; SuperKart Revenue Prediction

Strongest signal: 12+ years of senior analytics experience plus the ability to ship practical ML and GenAI workflows when the business problem calls for them.

Pinned Loading

  1. telecom-customer-churn-prediction telecom-customer-churn-prediction Public

    Description: Telecom customer churn prediction with calibrated ML, threshold tuning, and Streamlit retention analytics

    Jupyter Notebook

  2. ab-test-analyzer ab-test-analyzer Public

    Interactive A/B test analyzer for landing-page experiments with statistical inference, confidence intervals, power analysis, Bayesian checks, ROI planning, Streamlit app, and CI-tested notebook.

    Jupyter Notebook

  3. medical-qa-rag-assistant medical-qa-rag-assistant Public

    Inference-only medical Q&A RAG assistant with source-grounded answers, ChromaDB retrieval, citations, and safety guardrails.

    Python

  4. superkart-revenue-forecasting superkart-revenue-forecasting Public

    End-to-end retail revenue prediction project with XGBoost, SHAP explainability, prediction intervals, Streamlit app, FastAPI endpoint, tests, and deployment-ready artifacts.

    Jupyter Notebook

  5. hotel-booking-cancellation-prediction hotel-booking-cancellation-prediction Public

    Hotel booking cancellation risk scoring with XGBoost, Streamlit, SHAP explainability, cost-aware thresholding, and revenue-impact simulation.

    Jupyter Notebook

  6. AndreyLoskutov AndreyLoskutov Public

    GitHub profile README and data analytics portfolio for Andrey Loskutov.