Data Science Student at S.I.E.S College of Arts, Science and Commerce (Autonomous), Mumbai
Graduating: 2026

I'm a data science practitioner who builds end-to-end ML solutions — from exploratory analysis and model development to statistical validation and production deployment. My work spans time series forecasting, natural language processing, predictive analytics, and interactive data applications.

• 🔬 Research Areas: Deep Learning, Time Series, NLP, Sports Analytics
• 🛠️ Technical Focus: PyTorch, Feature Engineering, Model Optimization, Deployment
• 📊 Approach: Data-driven decision making with rigorous validation
• 🎯 Passion: Solving real-world Business problems with AI and statistical methods
• 💼 Status: Open to ML Engineering, Data Science, and Research opportunities

24-hour AQI predictions across 10 Indian cities using Temporal Fusion Transformer

• Built synthetic dataset from 99.7% missing data using Gaussian Process Regression
• Achieved 96.5% RMSE reduction through iterative model optimization (v1 → v3)
• Deployed interactive Streamlit dashboard with probabilistic forecasts (Q10/Q50/Q90)
• Tech: PyTorch, TFT Architecture, Scikit-learn, Streamlit, Plotly
• Results: R² = 0.9584, Statistically validated (p < 0.05)

📂 View Project | 📊 Live Demo

Multi-model NLP architecture for email security using BERT and LSTM

• Implemented dual-model approach combining contextual embeddings (BERT) and sequential patterns (LSTM)
• Engineered features from email headers, body text, and metadata
• Built robust preprocessing pipeline for handling diverse email formats
• Tech: BERT, LSTM, PyTorch, NLP, Scikit-learn
• Application: Cybersecurity, Email Filtering

📂 View Project

Sports analytics and match outcome forecasting using historical data

• Analyzed historical FIFA World Cup data with statistical modeling
• Developed predictive models for match outcomes and tournament progression
• Feature engineering from team statistics, player performance, and historical matchups
• Tech: Python, Scikit-learn, Pandas, Statistical Analysis
• Domain: Sports Analytics, Predictive Modeling

📂 View Project

• Frameworks: PyTorch, TensorFlow, Scikit-learn, Keras
• Architectures: Transformers (TFT, BERT), LSTM, CNN, Ensemble Methods
• Techniques: Time Series Forecasting, NLP, Feature Engineering, Model Optimization
• Specialization: Uncertainty Quantification, Multi-horizon Forecasting, Transfer Learning

• Analysis: Pandas, NumPy, Statistical Methods, Hypothesis Testing
• Visualization: Matplotlib, Plotly, Seaborn, Power BI
• Preprocessing: Feature Engineering, Data Wrangling, Synthetic Data Generation
• Statistical Tools: Gaussian Processes, Bootstrap Methods, A/B Testing

• Databases: MySQL, Oracle, Advanced DBMS
• Web Development: HTML, CSS, JavaScript, Bootstrap
• Mobile: Android Studio, Java (Beginner)

• Advanced ML: Transformers (GPT, Vision Transformers), Graph Neural Networks
• MLOps: Model Monitoring, CI/CD for ML, Experiment Tracking
• Big Data: Distributed Computing, Spark, Scalable ML Systems
• Specialized Topics: Reinforcement Learning, Federated Learning, AutoML

• Time Series Forecasting: Multi-horizon predictions, Uncertainty quantification
• Natural Language Processing: Transformers, Sentiment analysis, Text generation
• Environmental AI: Climate modeling, Pollution forecasting, Sustainability applications
• Sports Analytics: Predictive modeling, Performance optimization
• Explainable AI: Model interpretability, Feature importance, Trust in ML

I'm actively seeking opportunities where I can apply my data science and machine learning skills to solve impactful real-world problems.

Looking for:

• Data Science / ML Engineering roles
• Research collaborations
• Open-source contributions
• Kaggle competitions

📧 Email: kaustubh.n007@gmail.com
📄 Resume: View Resume
👨‍💻 Portfolio: All Projects

💡 "Data is the new oil, but insights are the refined fuel" 💡