Experience
Machine Learning Intern
Shopify
- Architected an E2E ML pipeline using XGBoost, dbt, and Airflow to predict conversion value for real-time ad auctions, increasing bid efficiency by 15% and generating $15M yearly in incremental revenue
- Engineered an online Multi-Armed Bandit (Thompson Sampling) framework for ad creative selection, solving the cold-start problem and boosting aggregate click-through rate by 25%
Machine Learning Research Intern
Valence Labs
- Designed active learning pipelines using uncertainty estimation to optimize data sampling efficiency, saving $40k in data acquisition costs for training in sparse-data regimes
- Engineered generative geometric deep learning models (GNNs) for 3D graph-structured data, implementing multi-objective optimization to satisfy 94% of complex structural constraints
Machine Learning Research Intern
Recursion
- Developed multi-modal foundation models fusing computer vision and high-dimensional tabular data; optimized transformer attention mechanisms to increase inference throughput by 4x
- Implemented domain adaptation and robust feature normalization techniques to handle distribution shifts across heterogeneous datasets, boosting out-of-distribution accuracy by 12%
Machine Learning Operations Intern
Nokia
- Engineered a MCP microservice to standardize the retrieval of proprietary data from an internal LLM store, reducing data preparation latency by 60% for domain specific training
- Built large-scale, Dockerized, Kubernetes-orchestrated microservices to source, validate, and register 100+ LLMs from HuggingFace into Nokia's MLFlow based model registry
Machine Learning Intern
Cisco
- Built an API integrating ChromaDB and an in-house LLM to convert natural language into structured Elasticsearch queries, adopted by 10+ Cisco engineering teams
- Designed and maintained ETL pipelines using Apache Airflow to ingest, transform, and load 5TB+ of data from internal APIs and cloud storage into a centralized Snowflake warehouse
Projects
Augmented Reality, zk-Proofs, Cohere, React, NextJS
Leveraged vector embeddings and similarity search to personalize geotagged AR scribbles; used zk-proofs for secure location-based access control
Python, scikit-learn, LightGBM, XGBoost, CatBoost, Optuna
Automated prediction system to identify market inefficiencies in ATP tennis odds, using a stacking ensemble model with custom time decay and ELO rating features
MLBets
Python, statsmodels, XGBoost, Poisson Regression, SQL
ML model using Poisson regression and gradient boosting to identify profitable pitcher strikeout props in MLB, achieving consistent edges (~$500/month passive income)
YOLOv8, Computer Vision, Medical AI
Brain Tumor Detection using YOLOv8 on MRI Scans — advanced computer vision for medical diagnosis using object detection
Spotify Web API, OpenAI GPT, Angular, Node.js, TypeScript
AI mood-based Spotify playlist generator that uses sentiment analysis to create personalized music experiences
Angular, Google Maps & Street View API, TypeScript
A free alternative to Geoguessr, hosted and playable; averaging over 50 monthly users
Python, Computer Vision, AI, Geolocation
AI-powered geolocation system that predicts locations from images using advanced computer vision and machine learning
Python, PyTorch, Reinforcement Learning, Deep Q-Learning
Deep Q-Learning agent using PyTorch that autonomously learns and optimizes strategies to achieve high scores in 2048
Education
Carleton University
Bachelor of Computer Science (Honours), Minor in Math
Expected: April 2026
GPA: 4.0Colonel By Secondary School
International Baccalaureate Diploma
Sep 2018 – Jun 2021
IB DiplomaSkills
Programming Languages
ML & Data Science
MLOps & DevOps
Cloud & API
Databases & Data Stores
Volunteering
Mentor
Technovation Girls
- Mentored a team that became global semifinalists, fostering skills in innovation and teamwork
- Guided young women in developing entrepreneurial skills and technological solutions for real-world problems