Nikita Pardeshi

• Automated infrastructure deployment with AWS CloudFormation, managed databases using RDS and integrated Chef for configuration management reducing deployment time by 60%.

• Set up and managed Amazon Redshift clusters and performed complex SQL querying for data analysis and reporting.

• Collaborated with cross-functional teams to provide product consumption and trend analysis contributing to a 10% cost reduction in operations.

• Developed a custom application in Python and Flask for data pipelines, significantly reducing data ingestion and cleaning efforts.

• Extracted, cleaned, and stored Chicago crime data with Python and GCP BigQuery SQL, reducing latency from 65 seconds to under 2 seconds.

• Developed data dashboards with Microsoft Looker and Plotly DASH, boosting decision-making for NGOs by 25%.

• Worked extensively with geospatial data and mapping in Python and Plotly.

• Implemented machine learning algorithms (AdaBoost, XGBoost, ElasticNet, Random Forest, SGD, SVM, ARIMA, SARIMA, Exponential Smoothing) in Python and R, with hyperparameter tuning.

• Provide users with Model interpretability and explainability features using InterpretML in Python.

• Developed Docker images for real-time model training and testing, deploying models on AWS with SageMaker, Lambda, and EC2.

• Built R&D dashboard for Reckitt – data collection, creating APIs in Flask and Django, database design and management, custom BERT models for automated categorization with a focus on News and Text Analytics (Natural Language Processing).

• Market sizing forecast models in Python, Social media Sentiment analysis.

• Hands-on experience with Atlassian Jira and Confluence for Agile Project Management.

• FMCG sector analysis (region and segment wise)- visualizations through insights, financial analysis in Python.

• Automated Company Earnings Call transcript’s collection, indexing and searching, summarization, text analytics and visualization. Lead a team of 10 data science interns.