Sindhu K

Data Scientist with 4+ years of experience in mortgage domain and data analytics. Skilled in Python, SQL, Machine Learning, Deep Learning, and Computer Vision using TensorFlow, PyTorch, and YOLO. Proven expertise in developing end-to-end AI solutions including real-time object detection, ETL pipelines, and interactive dashboards. Proficient in deploying models to production, analyzing business data, and building visual insights with Power BI and Streamlit.

• Data Visualization using Power BI, GUVI IITM Research Park
• Certified Data Science Professional, GUVI IITM Research Park

1. PhonePe Pulse Data Visualization Dashboard   

Tech: Python, SQL, PostgreSQL, Streamlit, Plotly
Summary:
Built an interactive data analytics dashboard using the PhonePe Pulse dataset with 13 tables of aggregated, map, and top-level data. Implemented dynamic filters for state, year, and quarter, created category-wise transaction insights, top 10 rankings, and choropleth maps. Integrated SQLAlchemy for backend queries and delivered a visually rich, real-time analytics experience.

Tech Stack: Power BI, DAX, Excel/CSV, Data Cleaning, Data Modeling

2.Luxury Housing Sales Analysis (Power BI Dashboard)

Summary:
Developed an interactive Power BI dashboard to analyze luxury housing sales across multiple locations and property types. The project focused on identifying key trends in pricing, demand patterns, and customer preferences. Performed data cleaning and built DAX measures to generate dynamic insights

 3. Content Monetization Model 

Tech Stack: Python, Pandas, Scikit-Learn, Machine Learning, Matplotlib/Seaborn

Summary:
Developed a machine learning model to predict YouTube video revenue using key performance metrics such as views, watch-time, likes, subscriber count, impressions, and CTR. Performed comprehensive data cleaning, feature engineering, and correlation analysis to identify the variables that impact creator earnings.

4. Amazon Music User Clustering

Tech: Python, K-Means, PCA, Pandas, Seaborn
Summary:
Applied clustering techniques to segment Amazon Music users based on their listening behavior. Preprocessed user activity data, reduced dimensions with PCA, and identified distinct customer groups. The clusters helped in understanding recommendation patterns and improving user personalization strategies

5. Fish Species Classification

Tech: Python, TensorFlow/Keras, CNN, Image Processing
Summary:
Built a Convolutional Neural Network (CNN) model to classify multiple fish species from image datasets. Performed data augmentation, image preprocessing, and fine-tuning of CNN architecture to improve accuracy. Delivered a high-performing classification model suitable for marine research and automated seafood processing.

6. Jio–Hotstar Advertisement Vision Analysis

Tech: Python, Computer Vision, OpenCV, ML
Summary:
Developed an ad-vision analytics system to detect, track, and measure advertisement visibility during live sports streaming. Used video processing techniques to identify logos, ad placements, and screen time. Generated insights such as ad duration, frequency, and brand exposure, supporting ROI measurement for advertisers.