Vishal MOHAN

Collaborated with architects and IT specialists to prepare requirements specifications and present the functional aspects and impacts of new features to stakeholders. Designed and implemented data pipelines for processing large-scale datasets using PySpark and distributed data processing capabilities. Worked on backend development using Python, designing scalable architectures and creating APIs. Managed servers on AWS EC2 Windows instances and implemented RDBMS/NoSQL/In-memory DBs for efficient data storage and retrieval. Developed Tableau dashboards with advanced features for tracking customer KPIs. Implemented money deduction solutions using Fastag with various providers. Conducted end-to-end testing, ensuring security and compliance, and utilized automated testing tools. Proficient in Linux-based operating systems. Implemented Jenkins for CI/CD pipelines and utilized Kubernetes for container orchestration, enabling automated build, test, and deployment processes. Collaborated with the DevOps team to optimize software delivery through efficient Jenkins and Kubernetes infrastructure.

Developed a PySpark project for analyzing large-scale datasets, leveraging distributed data processing capabilities. Implemented data transformations and manipulations using PySpark's DataFrame API, ensuring efficient cleaning, preprocessing, and transformation of raw data. Utilized PySpark's MLlib library for various machine learning tasks, including feature extraction, model training, and evaluation. Achieved a significant 50% reduction in processing time by designing a scalable data pipeline and implementing performance optimization techniques. Successfully integrated the PySpark project into a production environment, ensuring compatibility with existing systems and optimizing overall performance. Collaborated with a cross-functional team to build predictive models, such as logistic regression and random forests, achieving an impressive 85% accuracy for customer churn prediction. Employed technologies such as PySpark, Python, Spark DataFrame API, Spark MLlib, Apache Spark, and Hadoop for seamless execution.

Real time application based on twitter. Apache Kafka was used for data processing; data were received in the form of Tweet from twitter using Twitter API. Read the tweet feeds using Twitter Streaming API, process the feed, extract the hashtag, and send it to Kafka. Codes were written in Java and run-on Spark.

Project: Smart Camera. Tool: Darwin V7 Labs, used for annotation of data to create a dataset for training the model. Embedded system work on Rockchip Board on LINUX OS. Model – MobileNet-224 used for Image Classification on Python. Model – InceptionV3 used Image Prediction on Python and TensorFlow framework. Model – YOLOv5 used for Object Detection and DeepSort for Object Tracking on Python and PyTorch framework. Motion Detection on C++ using OpenCV Library. Model Training on GPU.