Andy Hendrickson Email & Phone Number

Minneapolis, Mn, Us

California, 94086, Us

3rd Party Data Pipeline:Took ownership of existing 3rd Party data pipeline, conducted process documentation gap analysis, developed SQL code enhancements, developed data extracts/views, conducted both data analysis and QC analysis and presented findings to business customers.SQL Server to Snowflake Data Pipeline Migration:Developed comprehensive migration plan, developed and tested migration code changes, coordinated data replication, developed additional surgeon referral data pipeline, coordinated user acceptance testing. SQL3rd Party Data Pre-Stage Pipeline Automation:Developed a Python process for automating a legacy time intensive 3rd party data preparation process incorporating raw text files and Web data inputs for Pre-Stage ETL. Dynamic Linear Programming Surgery Volume Estimation:Developed a script for estimating surgery volumes using Python, Pulp.Data Robot Administration:Served as Data Robot Administrator for helping new users obtain account access, and help train users on the machine learning modeling process used in Data Robot.

Bellevue, Washington, Us

Client Best Buy: Data Science & Optimization - Research & Development Team (June 2021 - Sep 2021)Working on ML Pipeline Implementation and Deployment on GCP.Client General Mills: Strategic Revenue Management (May 2020 - May 2021)Migrated on-prem Impala based model input data pipelines to GCP. Developed BigQuery tasks and configuration based DAG Airflow job workflow.Technology Presentations:-BERT Transformer re-model and fine tuning presentation.-Tensorflow customer complaint modeling. Tools: Python, Impala, GCP, Airflow, BigQuery, AI Platform / Vertex AI, Kubeflow, Seldon, BERT (Transformers), Keras, Tensorflow, sci-kit learn, PyTorch, SHAP, Bayesian Networks

Eagan, Mn, Us

Text Classification Models:Developed Deep Learning based models that automatically classify text input with a predicted category (and sub-category). Models used to automate high confidence text classifications for human expert classifiers.Image Classification Models:Developed Deep Learning based models that automatically detect the presence of surchargeable objects in input images.Actuarial Models:Updated past Actuarial models using machine learning methods.Bayesian Inference:Researched and developed prototype Bayesian Inference Multidimensional Item Response Theory models.Geospatial Model:Developed a model incorporating geospatial trilateration methods with natural language processing methods to determine the likelihood of property rental.

Bozeman, Montana, Us

-Data Mining Projects:Contributed scoring and segmentation models, research and development of key derived variables used as model input, exploratory data analysis, data preparation, model evaluation, and model implementation for multiple modeling projects for a Health Nutrition client. Tools: SAS, knime, python, R, Ab-Initio-Graph ETL Projects:Extensive experience in developing data processing graphs/flows using Ab Initio. Developed graphs that extracted input data from both file and database (Oracle, Netezza) sources, sorted and deduped data, applied complex business rules as transform logic in various components (Join, Dedup, Rollup, Scan, Reformat), implemented scoring model algorithms, and loaded processed data into both file and database outputs. -Check Redemption Fraud Detection Project:Developed report, for a Health Nutrition client, identifying addresses where potential product check redemption fraud occurred. The report could later drive business rules for automated assignment of addresses and customers to a fraud roster. Tools: Oracle, SQL, Ab-Initio, regular expressions-Marketing Data Error Identification Project:Developed a Data Warehouse process to automatically identify errors relating to marketing data for a Pharmaceutical client. Error combinations of media channel, survey, question, and answer data were aggregated and output. Tools: Netezza DB, Ab-Initio, SQL-Operations and Data Warehouse Post-Processing Lead:Led Operations and Post-Processing teams during transition from former Operations vendor. Trained new US and off-shore internal team members on data warehouse lifecycle. Conducted extensive gap analysis to replace legacy Operations process documentation and designed/developed automated process enhancements.-Automated Unica Marketing Campaign Scheduler: Developed a process to automatically schedule and run multiple IBM Unica sessions/campaigns in parallel. Tools: Unica database, Autosys, and Ab Initio.

Minneapolis, Mn, Us

-Heart Failure Prediction (2010):Partnered with a Research Scientist to develop a model to predict heart failure based on patient’s past data characteristics. Data Mining contributions: Data Understanding/Data Preparation, EDA, Variable Selection/Creation, Model investigation. Tools: SAS, PASW Modeler-Battery Failure Prediction (2009-2010):Partnered with researchers to both create a model to predict implantable device battery failure and a dashboard to Clinicians to coordinate device replacements based on predictions. Data Mining contributions: Data Understanding/Preparation, Variable Selection. Tools: SAS, Perl, Spotfire-Enhanced Episode Detection (2010):Partnered with researchers and U of M Machine Learning faculty to explore optimized per patient heart episode detection algorithms. Data Mining contributions: Preparing device data dictionaries, Data Understanding/Preparation, Researching Supervised and Unsupervised learning methods.

Minneapolis, Mn, Us

-Architecture/Data Modeling:Experience designing normalized/de-normalized database/data warehouse schemas. Designed centralized device data management system. Planning integrated data model from Clinical CRF, Device Registration, Device Network, Interrogated Device, and Episode Adjudication data sources. Data Modeling Tools: Sybase Power Designer, MS Visio.-Data Stewardship:Acted as Data Steward on a large cross business data integration project. Worked with other Data Warehouse groups to define a common device data vocabulary for data dimensions and business processes.-Data Extraction/Manipulation/Transformation (ETL):Extensive experience in using multiple tools(SAS, SQL, PL/SQL, Perl, PASW Modeler, XSL(T)) for extracting/manipulating/transforming data from File(XML, TXT, XLS), Web(Web Mining), and database(Oracle, SQL Server, Access) sources. -SQL Experience:Extensive experience writing complex SQL and XPATH/SQL statements for Database Views, XML DB, Reporting, Stored Procedures/Functions, and application support. Understand the Rewrite query form for accessing the XML Table. Also, created a new SQL/XPATH code tuning method which substantially improves query times for Oracle XML databases. -Procedural DB Programming (PL/SQL, SPROC):Extensive experience writing stored database procedures, functions, and packages in Oracle and SPROC stored procedures in SQL Server. -Automated XML Loading ProcessDesigned/Developed an automated XML loading process which loads XML Device Interrogation data into an Oracle XML DB, then appends current data into indexed static tables from XPATH/SQL views. A flag based control table and PL/SQL procedures, functions, packages were written in combination with a fully parameterized KORN Shell control script that features parallel command calls and script based rollbacks to prevent corrective work when errors occur.

Minneapolis, Mn, Us

-Centralized Device Data Management Project (2008-current):Formed collaborations between Software, Research, Data Warehouse, and Clinical groups to plan and develop my design for a new Centralized Device Data Management group. Led project strategy team by facilitating project meetings, developing strategic vision, and leading presentations to executive management to obtain sponsorship. This project will provide greatly enhanced Analytics capabilities for developing new medical device algorithms for patient therapy.-Automated XML Device Interrogation Loading ProjectDesigned an automated XML loading process and guide a group of Oracle DBAs and developers through implementation. The project loads XML Device Interrogation data into an Oracle XML DB, then appends current data into indexed static tables from XPATH/SQL views. A flag based control table and PL/SQL procedures, functions, packages were written in combination with a fully parameterized KORN Shell control script that features parallel command calls and script based rollbacks to prevent corrective work when errors occur.-New SQL/XPATH Tuning Method:Created a new SQL/XPATH code tuning method which substantially improves query times for Oracle XML databases. -WebEGM episode adjudication web application Project:Led application development team in support of WebEGM. This application compares the Physician and device interpretations of the episodes which enables the System Engineers to refine the algorithms that deliver treatment to Patients. Led requirements gathering, facilitated customer and developer meetings, recommended user interface composition, designed data model enhancements (SQL Server)

Minneapolis, Mn, Us

SAS Clinical Reports:Created Clinical reports and datasets using SAS data steps, proc import, proc sql and others. Reports took input from various data sources including Oracle tables, SQL Server tables, XLS and TXT files and SAS datasets. Report output varied from XLS, HTML, TXT, and .sas7bdat.-Automated Test Scripts:Created a SAS program for automating Clinical report test scripts from a System Requirements document. This reduced the time required to create thereport test scripts manually and automated portions of the testing as well.

Boulder, Co, Us

-Fleet management application:Contributed design and development towards FleetTeam fleet management web application. Co-designed Oracle database schema and ERD using Oracle Designer. Created data merging load scripts using PL/SQL. Developed application using Java 2EE, HTML, JavaScript, XML and SQL.

Durham, North Carolina, Us

-IVR to web conversion:Created pioneer prototype patient enrollment, diary entry, randomization, and drug distribution web application using ColdFusion. This project initiated the creation of a new department within Quintiles. -Clinical PL/SQL packages:Created PL/SQL packages, procedures, and functions for an Oracle database. These were used in coordination with an IVR (Interactive Voice Response) system to allow doctors to enroll, randomize patients, and receive on-demand fax reports.

Ms, Data Mining

Ba, Physics