Karan Vijayvargiya Email & Phone Number

Toronto, On, Ca

Toronto, Ontario, Ca

- Developed MLOps Process and published Whitepaper to establish best practices and moving Machine Learning models to Production from a Development Phase in Databricks. - Led the development of Fraud Detection Project to identify suspicious transactions - Card Present as well as E-Commerce and achieved the target SLA for flagging out the risky transactions. This helped the Risk Team to leverage the Machine Learning based approach and helped them gain efficiency by moving from Rule based approach to ML approach. -Executed and Automated the workflow in Databricks for Data Monetization project to share Quality data insights with customers with an efficient Monitoring system in place.- Designed PowerBI Dashboard for Technology team to report out KPIs and Metrics for various departments.

Bangalore, Karnataka, In

- Successfully executed Anomaly detection of System logs using K-means for clustering, PCA for visualization and Countvectorizer+Tf-idf for feature creation. - Built MLOps Pipeline Monitoring System with Azure Databricks, Datafactory and Logic App with MLflow to detect data and model drift and trigger alerts in the form of email notification to user to retrain model in case of model getting stale with time.- Managing Epics and project management in Azure Board, Version control system in Azure Repo.

Gurgaon, In

● Worked in the NLU Development team for deployment of Advanced Virtual Assistant/Chatbots across multiple regions and languages for Tax and Accounting products. ● Trained custom Word2Vec Embeddings to be used for Intent Detection Pipeline. ● Implemented RNN and Chargrid Approaches alongwith Tesseract OCR for VAT related field recognition from Customer Invoices in Azure Machine Learning platform. ● Fine-tuned LayoutLM model - BERT based model to extract information from Invoice pdfs and used the information to classify a line item as VAT reclaimable or not. ● Leveraged Azure ML experiments to submit and track different runs, Blob storage for data storage and Azure Container Instance for deploying models to production. ● Implemented MLOps guide for the projects from business requirement gathering till Model Operationalization. ● Used Bitbucket for Version Control system and Confluence for documentation. ● Used Miro for PI Planning and TwistLog for Docker Image Scans.

Benton Harbor, Michigan, Us

● Leveraged VGG-16 model as a baseline and tuned the model as per the project requirement using python with the TensorFlow framework and Keras API to classify fabric between Knitted and Woven structure to be used by the Laundry team and achieved a training accuracy of 98% and a validation accuracy of 97%. Recognized by Managers for "Star Performer" Award for this work.● Participated in Digital Hackathon conducted at Global Organisation Level and implemented the idea "Voice Based Error Fix in Washers" to come up with a conversational AI application using BERT for SQuAD and Google Dialog Flow to build upon the conversation with the app. ● Worked closely with business and engineering teams to encourage statistical best practices with respect to experimental design, data capture and data analysis. Worked independently in Data Preprocessing Techniques to make data useful for creating Machine Learning Models

Bangalore, Karnataka, In

● Coordinated simulation support activities at client location (Stuttgart, Germany) for 3 months and efficiently working with product development teams on projects and evolutions, driving engineering functionality and conducting design reviews meetings.● Executed python and linux scripts to automate daily repetitive tasks which helped in increasing team efficiency from 70% to 87%.

- Worked in NVH Domain and replaced Akusmod software with Hypermesh for acoustic cavity modeling. Helped in reducing costs associated with previous software. - Developed CAE method to estimate the amount of welding distortions using LS-DYNA using implicit analysis in Manufacturing simulations team.

Went through a training in Automobile servicing and maintenance which involved dismantling and assembling of Maruti 800 and 1000, Ambassdor and a 2 wheeler engine(Hero Honda CD100) and study of each and every component of automobiles. Diagnosis and sensor analysis of the vehicle was also covered.

Worked on IR sensors to develop a vision based robot meant to clean floors.

Worked in Checks department of the company- Study of various techniques used in monitoring Fasteners for desired sizes and dimensions before it is despatched to customer.Worked on project titled "Standardisation of annealing process"- Data analysis of annealing Temperatures at various stages to make the process faster and efficient.

Master'S Degree, Data Science

Msc Physics(Hons.) With B.E. Mechanical(Hons.), Physics And Mechanical

Inter, Science With Comp. Sc.