Nour Elba

<p>Created a SaaS side-project to help people find jobs using state of the art machine learning + generative AI.<br>
Hacked together an app to help people find jobs, serving thousands of customers all around the world<br>
Developed a proprietary ML model that can match users with jobs by analyzing over 100 statistical features<br>
Helps job seekers find jobs that are a good fit for them and helps employers receive resumes of applicants that are already vetted to fit the profile of the job description</p>

<p>Launched a data science research company, applying large-scale ML solutions for enterprise clients globally, leading a team of data engineers/scientists.<br>
Project: Predict if a tweet is going to be viral in the next 48 hours—within one hour of that tweet being posted<br>
Result: Stitched four neural network models, achieving 90% accuracy (sMAPE). Solution is being used by a large US-based media intelligence company every day (with customers like SpaceX, BofA, Subway)<br>
Project: Detect article relevancy for an entity based on a large corpus of labeled data (300k+ articles). Primary goal is to enable predictions for new clients which have no history/labels<br>
Result: Implemented transfer-learning techniques to transform the data to fit a spectral temporal graph neural network, achieving an 84% F1 score for existing entities and 75% on average for new/sparse-data entities</p>

<p>Built an end-to-end system for operational data integration and weekly purchase recommendation reports, correlated with Brent Crude Oil futures<br>
Led a team of 3 UCLA interns across the entire lifecycle of deploying a machine learning product/solution</p>

<p>Created a discounted cash flow model for rate applications and infrastructure spend assessment<br>
Automated key data initiatives, saving significant time and resources annually<br>
Initiated project to migrate the organization’s data strategy to the cloud, architecting an end-to-end modern data stack<br>
Pioneered a cultural shift in data usage organization-wide through a "Data Stewarding" system</p>