🎓 Second-Year at Barnard College of Columbia University | Economics & Computer Science
🧠 AI/ML Fellow at Cornell Tech’s Break Through Tech AI Program 📍 Based in New York City | Originally from Nepal
🔗 LinkedIn

I’m a student passionate about the intersection of AI, economics, and public impact—especially in the context of emerging markets. As a Break Through Tech AI Fellow, I’m working on real-world ML problems that center equity and accessibility.

Beyond AI, my experiences include conducting scientific research, launching data-driven marketing campaigns, and writing for national publications. I thrive at the intersection of technology, policy, and storytelling—and I’m always excited to learn something new.

📧 avapokharel@gmail.com | ap4678@barnard.edu
🔗 GitHub

Languages: Python, Java
ML & Data Science: Pandas, NumPy, scikit-learn, Seaborn
Tools: Jupyter, Google Colab, GitHub, Excel
Soft Skills: Technical Writing, Research, Public Speaking, Communication

Built a binary classifier to predict income levels based on demographic data using logistic regression, decision trees, and SVM.

• Tools: Python, Pandas, scikit-learn, Seaborn, Jupyter
• Result: Achieved ~82% accuracy after optimization.
• Includes dataset, Jupyter notebooks, visualizations, and full documentation.

Contains lab assignments and personal projects from a Machine Learning Foundations course exploring core ML techniques.

• Projects include:
  1. Income Classification Model: Logistic regression, decision trees, random forest; evaluated with accuracy, precision, recall, and log loss.
  2. Regression Model Comparison: Compared Linear Regression, Ridge Regression, and Random Forest Regressor using MSE and R².
  3. Neural Network Implementation: Built a feedforward neural network from scratch with NumPy; trained with backpropagation and gradient descent.
• Tools: Python, NumPy, Pandas, Matplotlib, Seaborn, scikit-learn, Jupyter Notebook.

Implemented a simulation of the classic casino game Video Poker. The program supports betting 1-5 tokens and evaluates hands based on standard poker rankings with corresponding payouts.

• Features: Fair deck shuffling, card replacement mechanics, hand evaluation including pairs, straights, flushes, full houses, and royal flushes.
• Tools: Java (classes for Card, Deck, Game, Player), command-line interface, comprehensive game logic, and test suite.
• Includes two versions of the game constructor for normal play and testing with specified hands.

• ✅ Elizabeth Gould Neff ’27 Scholarship
• 🧠 Dean’s List (all semesters)
• 🏆 "Top in Country" Award for 2 A-level Subjects (Psychology and English General Paper)

• Advocating for ethical AI, inclusive tech, and financial literacy
• Knitting and reading books

-->