David Kersey

About

I am a Machine Learning Engineer with 3+ years of experience building and delivering machine learning and analytic solutions to clients.

Prior to my current role as ML Engineer at cBEYONData, I worked as a Consultant at Deloitte primarily delivering data science solutions in fraud. I recieved my MSBA from William and Mary in 2020.

Prior to my data science career, I worked in agriculture and wildlife biology which took me to work in palces such as Africa, Alaska and Montana. I am a RPCV Zambia. I recieved my Bachelor's degree in Biology from St. Mary's College of Maryland.

Work

Over the past 3 years I have delivered as a lead developer on projects with significant impact, for example, preventing billions of dollars in fradulent payments from leaving a state benefits system, developing an internal threat ML & BI solution for one of the largest utility companies in the US and instituting a ML and RPA solution for various federal agencies to automate their auditing process. In that time I have learned to focus on engineering solutions that start simple with grounded truths in the client's domain/system and build in complexity, bringing the end user along with me on that journey.

I am currently working in Databricks, managing and improving a large ML pipeline for 4 federal agencies. On a day-to-day basis, I use Databricks, MLFlow, Python and SQL.

Additionally, I am leading efforts to build LLM applications for my company. We are developing in Langchain, Haystack, Docker and Elasticsearch.

Projects

Large Language Models

I am currently building out a portfolio of projects realted to LLMs. I am interested in automating the process of fine-tuning models to build RAG Q&A applications. Please visit my LLM Github repo to see example work.

MLOps

Additionally, I am interested in improving MLOps across my personal and team ML projects. I will be building example projects utilizing MLFlow, CI/CD and other best practices. Please visit my MLOps Github repo to view my progress.

Fraud Detection

Finally, I will demo a sample of interesting methodologies I used in my fraud detection work. Please visit my Fraud Detection Github repo to view my progress.

Contact

• Linkedin
• GitHub
• Email

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';