Work Experience

Hoosier Community Network - Data Analyst, Jun 2024 - Present

• Built 5+ ETL pipelines using AWS Glue to process XML, CSV, and JSON data from 70+ counties in Indiana, storing in Blob Storage.
• Ensured data integrity and reliability by validating REST APIs in Postman, achieving a 30% reduction in data errors.
• Leveraged Databricks and Redshift with SQL and Python for data handling, cleaning, and transformation, improving reporting efficiency by 20%.
• Developed Tableau dashboards using DAX and Power Query to visualize time-series and geospatial data with demographic filters, tracking 10+ KPIs for stakeholders.

Magna International, USA - Data Scientist, May 2023 - Aug 2023

• Led 5+ Python ETL data pipelines to retrieve, clean, and process time-series sensor data from APIs and Excel, storing 80+ GB in Amazon Timestream/InfluxDB, improving data consistency across 10+ sensors.
• Leveraged Databricks for data pre-processing, created interactive dashboards in Grafana with custom queries and alerts for real-time monitoring, enhancing anomaly detection by 20%.
• Classified chiller tank flow using data modeling techniques - logistic regression, decision trees, random forest, with data from 100+ injection molding machines in Databricks, resulting in 10% malfunction reduction.

Continental, India - Data Scientist, Aug 2021 - May 2022

• Collaborated on improving ADAS vehicle speed control by developing a PointNet++ classifier using radar sensor data, increasing responsiveness by 20%.
• Extracted point-cloud and JSON data from MongoDB into S3 through 10+ APIs, implementing data cleaning and handling techniques that reduced retrieval errors.
• Processed point-cloud data using Spark, combining 3D positions, relative speeds, and semantic segmentation classes, which improved data integration efficiency by 25%.
• Developed and maintained Tableau dashboards to monitor model performance and KPIs (Key Performance Indicators) for speed changes, elevated model stability by 10%.

Anubhooti Solutions, Pune, India - Machine Learning Engineer, Jan 2021 - May 2021

• Engineered an Automatic Speech Recognition (ASR) system for 100+ recordings in AWS S3, leveraging Librosa in python for audio pre-processing and MFCC feature extraction.
• Optimized the ASR Triphone model using transfer learning from a pre-trained Kaldi model and lexicon and vocabulary texts, achieving 90% accuracy through hyperparameter fine-tuning.