★ Working as Machine learning Developer in a Life Science Product which deals with Breast Cancer Detection using Deep Convolutional Models. Using Cutting edge Models such as EfficientNet, UNet, Mask R-CNN, we are trying to do tumor segmentation and predict the All-red Scores for multi-stained slides.★ Primary working as Module owner for Region of Interest (ROI) algorithms which tries to segment tumor regions from other regions such as normal, Folds, DCIS regions. This helps subsequent models to only process relevant regions given by ROI models.★ Customized EfficientNet Model & TensorFlow data Generator to support huge volume of data across multiple GPUs. Modified dice loss function and IOU metrics to ignore classes.★ Designed & Developed Preprocessing Framework and TensorFlow Training Pipeline to process large medical tiff slides with more than 80% GPU utilization throughout.

★ Working as a Lead Engineer in a product that deals with structured data contextualization and can perform data Normalization, Ingestion, harmonization, and Mapping in large scale. ★ I have worked closely with Engineering and Management in Transforming the Business into viable Product features which have yielded improved Customer Satisfaction. As a result, more than 5 POCs got converted to approx. 100 million Business.★ I was also responsible for the Research and development of techniques like lazy loading, batch processing that have boosted our existing accuracy and performance by 80%.★ The product extends the capabilities of using existing Machine Learning Algorithms like FastText, Naive Bayes to Custom Text Processing Algorithms that can be injected to the Product dynamically based on the business use cases.★ Designed and architected the custom Dataflow Framework using core Python multiprocessing libraries and Socket Programming is horizontally scalable, cloud computable (AWS) and uses action-event methodology. This design can process 100k records per hour and is up for patent evaluation.★ I have also Programmed Algorithms like Bag of Words, Word Graph, Abbreviations, scoring metrices using Sklearn, Pandas, NumPy etc which are being consumed by the product extensibility.

★ Created and implemented Flask Rest APIs which accelerated the product communications with other products.★ Containerized the modules using Docker which helped in decreasing the production deployment cycle from 1 day to couple of hours.★ Developed and Enhanced MongoDB queries for better performance in replicated and sharded clusters.

★ Worked as Software Engineer on multiple products like GoPro Camera Smart Video Application ,Smart Watch ,Continuous Delivery Platform using Android, NodeJs, Java, MongoDb. ★ Also have used AWS cloud for deployments and storage.

★ Digital Ads-QA Automation Frame-work ( Java ) ★ This is an Automation Project of entire Digital Ads QA Process. This processconsists of Creating test pages in DFP tool, taking Ads specifications using Charles tool & generating Charles Report.★ Android Application using Gopro camera and its SDK . This application can beused in various Sporty & adventures activities, showing Real time video Recordingwith some additional metadata overlayed on the video.(Augmented Reality).★ Android Mobile application for Refueling their vehicle in which user can providetheir current location Using Google Maps and the vendor will go to that location andrefuel the vehicle.