PRANAY REDDY DAVA

Practical Data Engineer possessing in-depth knowledge of data manipulation techniques and computer programming paired with expertise in integrating and implementing new software packages and new products into the system. Offering a 7-year background managing various aspects of development, design, and delivery of data solutions. Tech-savvy and independent professional bringing outstanding communication and organizational abilities.

Action Recognition using RNN(TENSORFLOW)

• Built an LSTM model for recognizing the action done in the video (streaming data).
• Used pre-trained Inception V3 (transfer learning) for feature extraction of sequential data from the videos.
• Achieved validation accuracy of 0.75 and top 2 categorical accuracy of 0.87

Convolution Neural Networks Architectures (TENSORFLOW)

• Implemented VGGNET-16, RESNET-18, Inception-V2.
• Implemented a total of 18 variants of the above architectures by changing activation (relu, Leakyrelu), regularization (Dropout, BatchNormlization), and optimizer (SGD, Adam).
• Observed the performance of these models on the CIFAR100 dataset.

Sparkify Data Warehouse {Python, Airflow, Spark, Amazon Redshift,-  S3}

Built a data pipeline using Apache Airflow and Spark to ingest data from a fictional music streaming app called Sparkify to Amazon Redshift data warehouse  which can be used to find insights into what songs users are listening to.