Akash Kanojiya

Built ETL pipelines to extract data from Hana Cloud to SAP DI and integrate it with Machine Learning models. Worked on data pipelines to collect data from multiple Hana tables and create master data using SQL for presentation to stakeholders on the Dashboard. Developed end-to-end machine learning prototypes and scaled them in production environments. Increased efficiency by 23%. Built LSTM Time series model to predict commodity prices with 16% higher accuracy than human estimation which helped the stakeholders to make the best procurement decisions. Web Scrapped large amounts of data, and performed data cleaning and preprocessing steps to feed in the existing model. Improved the model performance by 10%.

Built an automated Python tool using regular expressions to clean the financial HTML reports. Worked on improving the efficiency of the existing machine-learning models. Assisted in developing a closed-domain question-answering model using the Bert NLP technique.