Stephy J

Data Science professional with 2-3 years of experience in Machine Learning, Deep Learning, and Generative AI, including Agentic AI workflows. Skilled in MySQL, Power BI, and Excel for data analysis, with hands-on experience building and validating predictive models.

Jr. Data Scientist at Excelr Solutions (2025-01 – Present)

• Built and fine-tuned classification models using XGBoost, LightGBM, and neural networks, pushing prediction accuracy to 95% and lifting overall performance by 20%+ through careful feature engineering and hyperparameter tuning.
• Took ownership of the full model lifecycle — cleaning data, engineering features, evaluating performance, and deploying via Streamlit and Flask pipelines — so non-technical stakeholders could actually use the results.
• Tackled class imbalance with SMOTE while working across demographic, behavioral, and economic datasets for binary classification problems.
• Analyzed 40,000+ text records using NLP techniques like tokenization, lemmatization, and TF-IDF to build a fake-news detection system, testing multiple models to find the best balance of accuracy and speed.
• Used clustering methods (K-Means, hierarchical, DBSCAN) to group countries by socio-economic indicators, turning the results into data-backed policy recommendations.
• Prototyped generative AI tools with LangChain and Hugging Face for real business automation use cases, building hands-on experience with large language models along the way.
• Built a weather agent powered by real-time API data, using Retrieval-Augmented Generation (RAG) with a Qdrant vector database to deliver accurate, context-aware responses.
• Mentored 50+ students in data science fundamentals and project execution, while coordinating planning across team-based initiatives.

Data Associate at Pavizham Healthier Diet Pvt Ltd (2022-08 – 2024-01)

• Built and maintained performance dashboards that gave leadership a clear, up-to-date view of business profitability.
• Analyzed sales and cost data for an international rice trading initiative, producing revenue reports that informed pricing and business decisions.
• Partnered with the marketing team to translate raw transaction data into profit insights that supported growth strategy.
• Maintained accurate records through structured data entry and reporting in Excel, and supported stakeholder meetings with timely, reliable data.
• Used data analysis to flag process inefficiencies and recommend operational improvements.

Trainee Associate at Speridian Technologies (2020-12 – 2021-02)

• Built a Student Management System (PHP, MySQL, JavaScript, AJAX) to digitize and simplify student record handling.
• Designed database structures and added interactive features that improved system speed and usability.
• Gathered requirements directly from stakeholders to ensure the final system matched real business needs.
• Completed a parallel three-month program with Cloud CX, gaining early exposure to cloud platforms and professional project workflows.

Professional Data Science Certification in Data Science – ExcelR Training (2024-08) B.Tech in Computer Science Engineering – APJ Abdul Kalam Technological University (2020-10) 12th Grade in Computer Science – St. Anne's Girls H.S.S. (2016-05)