Data Scientist & Al Consultant with over 10 years of experience developing Al and data solutions, combined with leadership in team development, client consulting, and strategic implementations. Skilled in translating complex Al technologies (LLMS, NLP, computer vision) into practical business solutions. Strong background in guiding client projects, leading multidisciplinary teams, and contributing to organizational Al growth. R. T. is skilled in translating complex Al technologies (LLMS, NLP, computer vision) into practical business solutions, and has a strong background in guiding client projects, leading multidisciplinary teams, and contributing to organizational Al growth.
Migrated data workloads from R to Python, re-architected model pipelines, and delivered performance optimizations for improved efficiency and maintainability.
Designed and implemented production-grade data models using SQL and DBT, with a focus on modularity, lineage transparency, testability, and scalable transformation flows within the QCue pricing platform.
Streamlined pipelines via query plans, DBT strategies, caching, and validation, boosting reliability and lowering latency.
Increased data usage efficiency by 24% by developing scalable ML models for needs analysis and churn prevention.
Collaborated with domain experts (NGO workers, policymakers) to align solutions with future marketing.
Experimented with generative Al for marketing campaign support (text & customer segmentation).
Excel
PythonXGBoost/Sklearn
Generative Al
Explainable Al
Context Labs
Senior Data Scientist (freelance)
2023 - 2025 (2 years)
Remote
Cut data processing time by 30% via optimized ETL pipelines, built 20+ ETL/ELT pipelines (Python, Spark, dbt) for 10+ TB/day processing, improving analytics readiness by 40%.
Used Git version control and collaborated via feature branch workflow and code reviews, applied interpretable ML (NLP, graph analytics, geospatial data) for sustainability (CO2 accounting, supply chain traceability).
Developed generative Al use cases such as automated supply chain data summaries, reduced DBeaver/Trino infrastructure costs by 30%.
Designed end-to-end pipelines for multi-source datasets, improving efficiency by 40%, optimized NoSQL processes via partitioning and indexing, reducing latency by 50%.
Built ETL scripts for 10+ TB datasets, improving data quality for key decisions, designed and implemented automated data validation workflows using Azure Data Factory and Azure Databricks.
Applied NLP and large language models for multilingual report classification and translation, used dbt for modular transformations, reducing reporting errors by 25%.
Utilized NLP and deep learning techniques to design a Seq2Seq model for a chatbot, using TensorFlow and PyTorch frameworks. The bot was finely tuned to provide context-aware, human-like responses, emulating the tone and expertise of a real customer service representative.
The AI-powered chatbot provided more tailored support, resulting in a 25% increase in customer satisfaction, a 10% reduction in customer wait time, and a 7% improvement in Net Promoter Score (NPS).
The project showcased the potential of NLP, deep learning, and AI-powered automation in revolutionizing customer service, contributing to more personalized, scalable, and efficient customer interactions.
A robust churn prediction model was developed, which analyses subscribers' historical behavior to accurately forecast the likelihood of churn. This was initially done using logistic regression but was later enhanced using advanced data science and machine learning techniques.
Utilizing social network analysis (SNA) for insights into subscribers' interaction patterns, variables capturing the influence of network dynamics on churn behavior were created. The variables were integrated into an ensemble model combining XGBoost and Random Forest algorithms, improving the model's predictive power. Iterative optimization and hyperparameter tuning further boosted the model's accuracy by 18%, outperforming the baseline.
This project demonstrates the effective use of AI, machine learning, and data-driven decision-making in solving complex business problems, enabling proactive customer retention strategies, and reducing churn rates. SNA and ensemble techniques were used, emphasizing the importance of diverse data sources and advanced analytics for actionable insights.
The project aimed to increase customer retention in underserved neighborhoods with limited internet access through advanced data science and AI techniques.
A web scraping bot was designed and implemented to extract, analyze and predict competitors’ expansion plan trends and opportunities; this data fueled machine learning models for customer segmentation and churn prediction.
The AI-powered strategies resulted in a 40% increase in retention rates and provided a scalable framework for future data-driven decision-making.
Developed an AI-powered recommendation engine for a retail company, using machine learning algorithms to increase customer purchase volumes through real-time product suggestions.
The system was built with Python, Scikit-learn, and TensorFlow, incorporating customer segmentation and predictive analytics for personalized recommendations and future buying behavior anticipation.
The implementation resulted in a 15% increase in average order value, 20% boost in repeat purchases, and 10% reduction in cart abandonment, demonstrating the impact of AI in optimizing retail strategies and enhancing customer engagement.
Developed an AI-powered platform for the oil and gas sector that provides predictive maintenance, emissions management, and sustainability compliance. It integrates machine learning, IoT sensor data, and advanced analytics to deliver actionable insights for operational excellence and safety.
The system leverages forecasting models and anomaly detection algorithms to predict equipment failures and maintenance needs with 90% accuracy, reducing unplanned downtime by 20%. Integrated IoT sensors monitor greenhouse gas emissions in real-time, aiding in identifying emission hotspots, forecasting trends, and reducing the organization’s carbon footprint by 15%.
The platform incorporates automated reporting tools to streamline sustainability certification, saving hundreds of manual hours annually. It provides insights for improving sustainability metrics and aligns with global initiatives such as the OGCI and the Paris Agreement, thus positioning the organization as a leader in sustainable practices.
