Ajayeswar Peddyreddy Email & Phone Number

Bentonville, Arkansas, United States

Austin, Tx, Us

Working on real-time location systems (RTLS) using BLE data

Seattle, Wa, Us

Mentored by Amazon Machine Learning Solutions lab and USF data institute:• Designing and researching new optimization algorithms for accelerating large-scale deep learning training.• Focused on the sampling strategy including selective back-propagation and importance sampling, and the attention-layer-emphasized sampling strategy on transformer models.

Clark, New Jersey, Us

As a Data Scientist, I worked on the product delivery of an estimated annual revenue of ~3M$. I handled Machine Learning and Software Engineering tasks for the AI product:Invoice / Packaging Slip Reader: Reads images / PDFs to auto-populate forms.ML:• Developed an NLP pipeline for invoice processing leveraging Azure OCR. • Trained an ML model on Azure Customvision using few-shot learning to detect tables in an image.Software Engineering:• Decoupled a monolithic API into a microservices architecture, resulting in 7 independent microservices with improved scalability.• Reduction of over 90% in the error rate for processing complex invoices by revamping the application’s architecture. Migrated the REST API from Flask to FastAPI and transitioned from synchronous to asynchronous processing with increased fault tolerance and agility.• Deployed the WebApp using Docker and Azure Kubernetes in production. The system serves 10,000 requests per day.

Clark, New Jersey, Us

• Reverse Image Search: Used pre-trained Res-net 50 architecture to encode the images and used modified K-nearest neighbors to classify the images and recommend the appropriate item.• Reduced false-positive rate from 25% to 5% using classification models and utilized YOLO for accurate object detection, to identify multiple objects from one image and recommend accordingly.• Developed a data pipeline using Azure Data Factory (ADF) to optimize data transfer across various CosmosDB in different regions, automating the process with scheduled triggers.

Master Of Science In Data Science

Bachelor Of Engineering - Be, Electrical And Electronics Engineering