Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Data scientist and NLP/Time-Series focused professional with strong end-to-end experience in data collection, cleaning, preprocessing, feature engineering, model evaluation, and deployment. Skilled in Python/SQL, Power BI/Tableau, and modern ML/DL frameworks to deliver actionable, results-driven insights. I build forecasting and predictive systems (including temporal fusion transformers and LLM-based approaches), validate models rigorously with time-aware methods, and translate complex datasets into business strategy using dashboards, reports, and clear presentations.…Data scientist and NLP/Time-Series focused professional with strong end-to-end experience in data collection, cleaning, preprocessing, feature engineering, model evaluation, and deployment. Skilled in Python/SQL, Power BI/Tableau, and modern ML/DL frameworks to deliver actionable, results-driven insights. I build forecasting and predictive systems (including temporal fusion transformers and LLM-based approaches), validate models rigorously with time-aware methods, and translate complex datasets into business strategy using dashboards, reports, and clear presentations.

Kaushtubh Narayan Kar

Data Scientist, Data Visualizer, Data Analyst, +3





Data scientist and NLP/Time-Series focused professional with strong end-to-end experience in data collection, cleaning, preprocessing, feature engineering, model evaluation, and deployment. Skilled in Python/SQL, Power BI/Tableau, and modern ML/DL frameworks to deliver actionable, results-driven insights. I build forecasting and predictive systems (including temporal fusion transformers and LLM-based approaches), validate models rigorously with time-aware methods, and translate complex datasets into business strategy using dashboards, reports, and clear presentations.…Data scientist and NLP/Time-Series focused professional with strong end-to-end experience in data collection, cleaning, preprocessing, feature engineering, model evaluation, and deployment. Skilled in Python/SQL, Power BI/Tableau, and modern ML/DL frameworks to deliver actionable, results-driven insights. I build forecasting and predictive systems (including temporal fusion transformers and LLM-based approaches), validate models rigorously with time-aware methods, and translate complex datasets into business strategy using dashboards, reports, and clear presentations.

Available to hire

Data scientist and NLP/Time-Series focused professional with strong end-to-end experience in data collection, cleaning, preprocessing, feature engineering, model evaluation, and deployment. Skilled in Python/SQL, Power BI/Tableau, and modern ML/DL frameworks to deliver actionable, results-driven insights.

I build forecasting and predictive systems (including temporal fusion transformers and LLM-based approaches), validate models rigorously with time-aware methods, and translate complex datasets into business strategy using dashboards, reports, and clear presentations.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Language

English

Fluent

Marathi (Marāṭhī)

Fluent

Hindi

Fluent

German

Beginner

Work Experience

Data Scientist Intern at Prodygi InfoTech

February 1, 2024 - April 30, 2024

Conducted end-to-end data workflows on 500k+ records using EDA, data cleaning, and transformation to improve decision accuracy by ~15% and reduce processing time by ~30%. Enhanced model performance by ~12% through feature engineering and statistical analysis supporting key business strategies. Built and applied predictive modeling pipelines to derive actionable insights.

ML Engineer (Phishing Email Detection) at Independent Project

January 1, 0000 - January 1, 0000

Engineered an end-to-end phishing email detection system using ML/DL with 96% validation accuracy. Implemented TF-IDF and embedding-based preprocessing and fine-tuned BERT as the top-performing model across evaluation metrics. Deployed as a real-time Streamlit web application for interactive email classification.

Data/ML Engineer (Temporal Fusion Transformer for AQI) at Independent Project

January 1, 0000 - January 1, 0000

Performed multivariate time-series analysis and forecasting for 24-hour AQI predictions using a temporal fusion transformer. Achieved R²=0.85 and RMSE=20.3, improving accuracy by 10–15% versus baseline models (LSTM, GRU, XGBoost) using feature engineering and temporal validation.

Researcher / Author (AQI Forecasting) at Zeno-do (Open Access Repository)

January 1, 0000 - January 1, 2026

Presented a temporal fusion transformer-based multivariate time-series AQI forecasting approach across multiple cities. Delivered AQI prediction with deep learning time-series methods and published in 2026 (open access).

Data/ML Projects (FIFA World Cup Outcome Prediction) at Independent Project

January 1, 0000 - January 1, 0000

Built an end-to-end match-outcome prediction model using historical FIFA World Cup data. Applied EDA and feature engineering to improve reliability and achieved ~95% accuracy. Implemented in Python with Pandas/NumPy and ML for predictive modeling.