Working on language model training, prompt engineering, and coding.

• Data Science: Working on census machine learning project based on ongoing needs. Performing text mining with Python, R based on open source embeddings. Developing ETL multi hop pipelines filter, aggregate, and modify with user defined functions various data sources and data streams. Design and update model features based on these discoveries. Keep track of model catalogues to apply relevant models to different parts of the project and maintain a record of a series of updated models. Designing, modifying, training, and updating models in Python and R for better customer segmentation for various situations as well as accommodating to different model stages. • Data Engineering: Performing SSIS and ETL operations to migrate data from various sources to SQL server for data integration, management, data mining, text mining, and data modeling. Deploying and monitoring updated models to AWS. Designing, expanding, and updating current business logics.• AWS and GitHub: Maintaining the GitHub branch for census processing, deployed the updated project to AWS and routine relevant AWS management.

• Data Science and Data Engineering: Creating Scope scripts & supporting C# extensions to collect metrics for map data of different sources; developed ETL pipelines and workflow solutions to transfer data between different databases; writing java files to collect metrics for open street map data. Utilize Oracle 19C Oracle Spatial to ingest various geospatial data types. Performing data mining and text mining query, and extract terms for analysis and visualization. Using ODI to automate data integration, ingest, and data security, and data quality check including loading data to SQL DB with parallel processing with SSIS.• Data Analysis: Creating BI Tabular data model on Azure analysis services server. Creating a set of measures and KPIs to generate reports based on the data model. The model is connected to Power BI Dashboard and then queried from there for presentation purposes. Also, the key performance indicators (KPIs) are important indicator for guiding and evaluating operation and marketing strategies. • Scripting: Developing java classes that examine geometry of atlas files from open street map raw pbf.

• Machine learning and deep learning: Creating NLP machine learning models to analyze and classify Azure failover cluster log files to detect anomalies and predict anomaly types. Based on customer analysis and segmentation, we identified types of anomalies to detect, which defined how to collect as well as process data from engineers.• Cloud Computing, REST APT, Python, C#: Implementing the whole application with Rest API, C#, and Python to automate the whole process from MS SQL server to web presentation cooperating with team members. Both data storage and models are deployed to Azure.• A/B Testing Experimental Design: Designing different sets of key features to describe the data sets and created A/B testing experiments to quickly collect insights and feedback to improve data models, assist machine learning processes, and guide the overall progress of the project. • Scripting: Writing SQL server scripts and stored procedures to perform various data driven tasks; creating C# code for Azure functions of web based services; developing python scripts and pipelines for machine leaning models.

• Data conflation with GeoProcessing and spatial analysis: Analyzing & comparing geometry and attributes of two commercial map databases for road overlapping analysis and map data merging. • Scripting: Writing C# and Cosmos scripts for the above analyses. • Outcome: The methods have been tested/evaluated for major markets (USA, UK, India, etc.)