Caitlin Ye

United States

@carnspiger

Machine Learning Engineer at Citi

Badges

Problem Solving
Sql

Certifications

carnspiger has not earned any certificates yet.

Work Experience

  • Machine Learning Engineer

    Citigroup Inc.•  July 2017 - Present

    ❑ Built Docker images for automated installs of analytical tools such as Anaconda Enterprise Notebook, Tensorflow/Keras on GPU, Datarobot, and H2O Steam for use as BlueData sandbox environments. Used Kubernetes for Anaconda Enterprise Notebook 5 orchestration. ❑ Engineering products for GPU such as H2O Driverless AI, Tensorflow, and Keras. Conducted testing of GPU load and memory using Tensorflow’s GPU functionality using a CNN with CIFAR-10 image dataset on Nvidia Cuda 9.0 systems. ❑ Daily experience setting up and utilizing Cloudera Hadoop services such as HDFS, Yarn, Hive, Spark, and Impala with Cloudera Manager. Experience with MapReduce jobs and Hadoop Linux commands. Experience building, distributing, and activating parcels for Cloudera Manager. ❑ Connects to Hive Server 2 and Spark from a custom Anaconda distribution I built using R (sparklyr) and Python (Pyspark/Impyla) for ETL processes. ❑ Extensive experience setting up and configuring Nginx as a web server, load balancer, and reverse proxy while meeting Citi’s strict security guidelines. ❑ Natural language processing using NLTK for analyzing internal ticketing system records. Utilized VADER library for sentiment analysis of social media data. ❑ Graph analysis for entity resolution with Neo4j. Analyzing the data allowed us to make connections for who is the same person, who knows each other, and their relationship. ❑ Feature engineering with DataRobot. Utilized this tool to train and model large data sets. ❑ Familiar with security verification processes of large financial institutions and how to avoid security flaws by implementing various authentication methods including SAML, LDAP, PAM, and Kerberos. ❑ Managing Mongo and Postgres databases in the backend of various analytical tools. Experience using Mongo as a Service. ❑ Automated analytical tools installs and environment configuration processes by utilizing Chef cookbooks. ❑ Wrote test automation scripts in Java using Selenium. ❑ Trained all new additions to our team.

  • Freelance Data Engineer/Data Scientist

    PumpJack Dataworks•  January 2019 - April 2019

    ❑ Provided preliminary design for Big Data Analysis Pipeline with batch processing and data streaming components. ❑ Delivered high-value business artifacts for use by client to assist in raising funds. It took them from Series B to Series C funding for their startup. ❑ Currently working on their data accessibility problem. Finalizing plans with new data size quote which requires compatibility for use with various data science and machine learning techniques which sets the company apart in their field of business. ❑ Now knowledgeable about SOC 2 and PCI compliance when using AWS. Also, have experience with PCI compliance architecture from working at Kibo on the OMS (order management system). ❖ Technologies Used: SOC 2/PCI compliance, Amazon Web Services - EMR, Embulk, Airflow, S3, RDS, Kafka, Spark Streaming, Impyla, Lambda, DynamoDB, Fluentd, Neo4j

  • Technical Consultant

    Kibo Commerce•  February 2017 - July 2017

    ❑ Works directly with clients, providing solutions during the software implementation phase and offers insight on configuring, customizing, and integrating with the Kibo eCommerce OMS platform. ❑ Create tools that stream RESTful API data for internal use in Java. ❑ Coaches client on implementing Kibo JSON and XML APIs based on RESTful API architecture. ❑ Troubleshoots client OMS issues utilizing Postman and JSON/XML API calls to help integrate client systems. ❑ Utilize SQL commands to search OMS databases to troubleshoot client orders and account information.

Education

  • University of Texas at Dallas, Richardson

    Software Engineering, BS•  January 2013 - May 2017

    Specialization in Artificial Intelligence.

Skills

carnspiger has not updated skills details yet.