Andy Yang Email & Phone Number

New York City Metropolitan Area

Early Promoted to Associate Level

New York City Metropolitan Area

• Continuing work from Rotation 2 as a Machine Learning Engineer, in addition to the below responsibilities.Project: Trade Surveillance System
• Creating a machine learning trade surveillance system to augment the current threshold-based trade alert system by analyzing features such as trade size and trade volume to save around $100k in compliance analysts costs.
• Built a logistic regression model pipeline and tuned logistic regression model to achieve a 96-97% accuracy score in detecting potentially fraudulent trades in foreign exchange and equities markets.
• Creating a list of potentially fraudulent trades and separating the trades into low, medium, and high-risk levels utilizing the probability scores created by the logistic regression model to increase operational.
• Utilizing Python to create a voice capture system consisting of three components: a recording program for microphone input, a preprocessing program to filter silent voice files, and a transcription program using OpenAI.
• Developed a virtual data assistant utilizing an in-house large language model to convert user inputs into SQL queries through prompt engineering techniques and few-shot learning.

New York City Metropolitan Area

• Project: Client Analytics Platform
• Utilized Pyspark to create jobs on Hadoop Clusters to process, filter, and transform internal email data.
• Developed unit tests for Pyspark jobs with Pytest and Mock libraries, testing the code quality and functionality of the data pipelines.
• Performed topic modeling of email text data using BERTopic and Latent Dirichlet Allocation (LDA) analysis, extracting the main topics and subtopics from the email content, and creating interactive dashboards to.
• Fine-Tuning Logistic Regression Model utilizing the Low-Rank Adaption of Large Language Model technique.Project: Large Language Models Research (Compliance Technology)
• Utilizing PyTorch to build and train deep learning natural language processing models (e.g., GPT) on a large corpus of compliance documents to create an intelligent virtual assistant which answers compliance-related.

New York City Metropolitan Area

• Refactored legacy Standard Instrument Protocol (SIP) data flows to a single centralized reference database to streamline regulatory workflows and increase operations efficiency.
• Implemented two-way SSL authentication within the codebase, replacing previous insecure basic authentication protocol.
• Backend development done through Python, built an internal tool called Know-Your-Data (KYD) which checks data quality (e.g., missing/conflicting data points) by comparing two JSON objects.
• Completed Deutsche Bank’s Google Cloud Associate Engineer Training Sprint Program, graduating with Distinction.
• Won the NYC Regional Hackathon – 2022 Global Technology, Data, and Innovation (TDI) ‘Transforming Mental Health.’ Created facial recognition software with Python to determine a patient’s emotional state. Consisted of.

New Jersey, United States

• Assisted Dr. Weihong Guo with process monitoring and statistical quality control to predict the quality of 3-D printed parts.
• Collected and wrangled over 200+ million 3-D printing data points and parameters stored in over 300.hdf5 files containing layer data of a 3-D printed part.
• Generated predictive models in Jupyter Notebook using machine learning algorithms such as k-nearest neighbors and support vector machines and compared the confusion matrix results to compare the different algorithms’.
• Created a binary classification model using a neural network algorithm and generated graphs showing model loss and accuracy vs. the number of epochs using Keras, sklearn, and matplotlib.
• Adjusted parameters such as the learning rate and batch sizes to see how the model loss and accuracy were affected and achieved a 96-97% accuracy rate.

New Jersey, United States

New York City Metropolitan Area

• Utilized Python (Flask) to build a central application connecting a combination of microservices architectures and MySQL databases containing internal employee and client data and processing unstructured data to.
• Tackled ad hoc feature requests through bi-weekly agile sprints and weekly code releases and utilized Fabric to deploy the Flask application to external servers automatically.
• Demonstrated the code function to various team leads based in the United States and London and onboarded incoming technology analyst graduates regarding application documentation and process.

New Jersey, United States

Master'S Degree, Information And Computer Technology

Bachelor Of Science - Bs, Mechanical Engineering (Energy Concentration), Magna Cum Laude

Bachelor Of Arts - Ba (Dual Degree), Philosophy