GOKUL N

Data Scientist with hands-on experience in Python, machine learning, statistical modeling, and NLP. Skilled at transforming complex, raw datasets into actionable insights that drive strategic decision-making and operational improvements. Passionate about developing innovative AI solutions to solve real-world problems, enhance business outcomes, and accelerate digital transformation.

1. Optimization of bus ticketing demand and forecasting

Client: a leading bus transportation company

• Developed machine learning models (ARIMA, Random Forest, Gradient Boosting) to forecast future bus ticket demand using historical sales data (~14,000 rows).
• Performed EDA, missing value imputation, and outlier treatment using Pandas and NumPy.
• Deployed the best-performing ARIMA model using Streamlit for real-time use by transportation stakeholders.
• Achieved forecasting performance with RMSE ≈ 112.5 and MAPE ≈ 40%.

 2. Automated transcript analysis to optimize recruitment decisions

 Client: One of the leading HR-tech companies that provides end-to-end recruitment solutions for top enterprises globally. They conduct thousands of interviews weekly across different geographies and roles, leading to high manual effort in evaluating and shortlisting candidates.

• Built a deep learning pipeline to classify interview transcripts (Fit vs. Not-Fit) using models like TextCNN, BiLSTM with Attention, and ELECTRA.
• Engineered features like sentiment, red flags, and skill count from unstructured text data.
• Achieved ~74.8% accuracy with TextCNN; deployed using Streamlit for real-time HR decision support.
• Addressed challenges like label imbalance, text noise, and metadata sparsity.

Tools & Technologies Used:

Python, Pandas, NumPy, Scikit-learn, PyTorch, Hugging Face Transformers, ARIMA, Streamlit, Matplotlib, MongoDB

• Certificate Program on Data Science, 360DigiTMG, 2024
• Career Essentials in Data Analysis by Microsoft and LinkedIn, Nov 10, 2024