ML / AI Analyst · Chicago, IL
Where
Intelligence
Blooms.
M.S. Business Analytics student at DePaul University. Former Research Associate at the Indian School of Business. IEEE paper forthcoming. I turned my own Instagram into a research dataset. The algorithm's most interesting subject.
Class of 2026
M.S. Business Analytics
DePaul University
8M+
reel impressions
own dataset
8M+
Reel Impressions
Own research dataset
92%
Model Accuracy
Customer churn prediction
6+
Years in Data
Finance → Research → ML
DePaul
Graduate Student
M.S. Business Analytics
Selected Projects
The Work
Chicago Closure Radar
ML early-warning system that flags Chicago restaurants and cafés at risk of closing months before they do. 312K+ city inspection records, XGBoost + SHAP, ROC-AUC 0.807. HIGH risk businesses close at 46× the baseline rate. Live app.
Customer Churn Prediction
Built a churn prediction model using XGBoost and Random Forest. 92% accuracy across 50K+ customer records. Added SHAP so the team could actually understand why the model flagged someone, not just that it did. Estimated 18% churn reduction.
Influencer ROI & Fake Engagement
I used my own Instagram as the dataset. 4,823 posts, 8M+ impressions, anomaly detection and NLP to find what was real versus what was bot traffic. Submitting to IEEE. Brands spend billions on influencer deals using gut feel. This gives them something better.
Admissions Analytics — DePaul Law
The admissions team was spending 20+ hours a week on manual Excel reporting. I automated it. Built forecasting models they now use to decide who to recruit and when to push harder on yield.
The Differentiator
The analyst who is the experiment.
Most ML analysts study other people's data. I built my own dataset from scratch by becoming a content creator. Three years of posting to Instagram, 4,823 posts and 8M+ impressions, became the dataset for IEEE paper on fake engagement and influencer ROI.
Before that: two years at the Indian School of Business doing real research. Before that: financial modeling at Whirlpool. Before that: supply chain at FedEx. Operations, finance, research, ML. That breadth is the edge most data scientists don't have.
Self as Dataset
Used @aneela_veldi as the research subject. Every data point was generated by real decisions, in real time, over 3 years.
Rigorous Methods
Anomaly detection, NLP, time-series modeling on data I generated myself. Then got it peer-reviewed.
Business Translation
Years of financial modeling means I can explain what the model means in dollars. That's rarer than it should be.
Interactive
Real or Bot?
Audit 8 Instagram profiles and spot the fake engagement. This is exactly what I built for my IEEE paper — at scale.
Case 1 / 8
0 correct
@wanderlust.daily
Travel89.3K
followers
16.3%
engagement
Spike
pattern
Engagement vs. industry avg (3.2%)
+42K followers in 5 days
Sample comments
Amazing shot! 🔥
Love this! ❤️
So beautiful! 😍
Gorgeous! ✨
97% single-emoji or 2-word comments
Recommendations
What colleagues say
“Aneela is one of the sharpest people I have worked with. She explains what the models are actually finding — and she makes people care about the results.”
Dr. Priya Ramesh — Research Faculty, ISB
“Her customer churn pipeline is still running in production. The SHAP explainability layer she added was what convinced leadership to act on the model's recommendations.”
Jordan K. — Data Engineering Lead
“What separates Aneela from other analysts is her finance background. She talks about ML results in terms of dollars and business outcomes. That is what actually gets decisions made.”
Prof. Michael Tan — DePaul University
“She's the rare analyst who also understands how audiences actually behave. That dual perspective makes her marketing analytics work far sharper than anyone else on the team.”
Sarah L. — Brand Partnerships Director
“The admissions dashboards Aneela built saved our team 20+ hours every reporting cycle. More importantly, they showed us what to actually do — not just what was happening.”
Enrollment Office — DePaul College of Law
“The Morningstar classifier project showed a level of financial domain integration that is rare in a graduate student. She knows when the model needs the business context, not just more data.”
Prof. Amanda Chen — Business Analytics
Let's Work Together
Ready to build
something remarkable?
Predictive model, data project, research collab, brand partnership. Whatever it is, let's talk.