ProgramsGraduate Consulting Experience (Summer 2024)ProjectsUsing LLMs to improve predictive AI for sales forecasts.

Project Title: Using LLMs to improve predictive AI for sales forecasts.

USAM Group Inc.

Details
Project Title Using LLMs to improve predictive AI for sales forecasts.
Project Topics Data Management Entrepreneurship Growth Strategy Reporting, Financial Planning & Analysis Research & Development Sales & Business Development Software Design & Development
Skills & Expertise Data Processing Data Visualization Database Administration Deep Learning Experimentation & Testing Instructional Design Quantitative Research Technical
Project Synopsis: Challenge/Opportunity
Work with our Chief Data Scientist to create sentiment analysis on natural language sales notes, feeding the results to an existing predictive AI model to improve its accuracy.
Project Synopsis: Activities/Actions Required
  1. Text Preprocessing:
    • Tokenize the text data into words or phrases.
    • Remove stop words, punctuation, and special characters.
    • Perform stemming or lemmatization to normalize the text.
  2. Sentiment Analysis:
    • Utilize natural language processing (NLP) techniques to perform sentiment analysis on the sales notes.
    • Apply sentiment lexicons, machine learning models, or deep learning models to classify the sentiment of each sales note (positive, negative, neutral).
    • Experiment with different sentiment analysis approaches and algorithms to determine the most effective method for the dataset.
  3. Integration with Predictive AI Model:
    • Understand the existing predictive AI model and its input requirements.
    • Modify the input features of the predictive AI model to include sentiment features derived from the sales notes.
    • Select the proper matrics for model accuracy.
  4. Integration with Production Environment:
    • Prepare the trained model for deployment in a production environment.
    • Integrate the model with the existing infrastructure, ensuring compatibility and scalability.
    • Test the integrated system to verify functionality and performance.
  5. Documentation and Reporting:
    • Document the entire process, including data preprocessing steps, model development, and integration efforts.
    • Create comprehensive documentation to facilitate knowledge transfer and future maintenance.
    • Prepare a report summarizing the methodology, results, and insights gained from the project.
Project Synopsis: Expected Results
Deliver a 1.0 release of code that scores sales notes across several factors. The desired delivery would include improved predictive accuracy, scalability and sustainability.

Project Timeline

Touchpoints & Assignments Date Type

Applications Closed for Students

 May 03 2024 Event

Students Upload Signed "Fordham Unpaid Internship Agreement"

 May 10 2024 US/Eastern (UTC-04:00) Event

Teams Finalized, Projects Assigned

 May 10 2024 Event

Industry Partners to Provide Each Offer Letter to Each Student

 May 17 2024 Event

Students Upload Resume

 May 17 2024 Action Item

Kickoff Eval

 May 24 2024 US/Eastern (UTC-04:00) Evaluation

Goal Date for CPT Approval

 May 31 2024 Event

Projects Launch!

 Jun 03 2024 Event

Temp Check

 Jun 12 2024 Evaluation

Temp Check

 Jun 28 2024 Evaluation

Temp Check

 Jul 10 2024 Evaluation

Projects End

 Jul 26 2024 Event

End of Project Self Reflection

 Jul 26 2024, 12:00 PM Evaluation

Program Managers

Name Organization
Beatriz Picard Fordham University