Responsible for dynamic data collection from Amazon Ad server, processing and storing in S3 bucket. Building and maintaining ML models and project pipelines such as time series modelling (e.g. Budget Shifting). Work extensively with Python, Sklearn, TensorFlow, Plotly, Docker, Apache Airflow and Linux/ Windows os for prototyping and model building. Quick prototyping of MVP and exploratory analysis. Working knowledge with Amazon advertising API and postman Carry out regular code reviews with the team lead. CONTRIBUTION: Lead and built an in-house dynamic data collection tool with the goal to automate budget shift (Amazon DSP) and Acos optimization (PPC) thereby replacing the existing 3rd party tool.

Technologies/Tools: Python 3.7, TensorFlow, Power BI, Azure Studio & Blob, AWS Sagemaker & S3 Machine learning-based deal size prediction model (Accuracy: 90%) Built and deployed machine learning-based decision-making system on the Azure cloud platform that will be used by 500 Diebold's global leadership (F1~94%) Created 5+ Dashboards on NPX and presented analytics insights to relevant clients that increase yearly revenue between 2 to 4 % Developed and presented 3+ Dashboards on Power BI (Win-loss analysis) Research Project: Touchless ATM, Touchless self check out

Technologies/Tools : Python 3.7, TensorFlow, NumPy, XML, Notebook 6.0, PyCharm Built an innovative GDPR General data protection rule compliant recommender system, which can safely influence 1.5 trillion US $ worth of business click Worked with big data sets 2800k transactions from the retail domain Use of Knowledge Graph embedding algorithms make the solution completely 100% scalable for similar data sets Optimized the efficiency of an LSTM by reducing memory consumption and processing time up to 50% and 85% compared to the initial version

Technologies/Tools: Autoit, Python 3.7, Ms Excel 2013, Matplotlib, SQL server 2016 Exploratory data analysis to identify the performance bottleneck of large (10 million lines of code) pos systems and optimize their performance (Best achievement: 300% processing time improvement) Partially automated the data pipeline containing thousands of rows of data which reduced manual workload by 80% monthly

Technologies/Tools: Appian 7.2/7.4, SQL server Developed a mobile tower location tracking application worth 1.5 million US $

Technologies/Tools : Core Java, Oracle database 12 c, Eclipse Kepler, Autoit Part of a 15 members team that implemented an infrastructure application used by .5 million health professionals worldwide Automated periodic database refreshment which increases the process efficiency by 66% and reduced manual workload by 10% yearly