TYLER EVANS
*****.*****.*******@*****.***
***Experience***
Laivly, Winnipeg, MB
Junior Machine Learning Engineer ML Team May 2019 – Present
• Reduced number of NLP models per tenant from 30 to 2; Retained F1 score by designing a custom architecture
• Enabled team of 10 to deploy models in production by designing serving code, preprocessing & model interfaces
• Reduced code duplication by proposing & developing features for a package used to bring experiments into production
• Eliminated model deprecation by introducing Docker containers to make experiments reproducible and scalable
Laivly, Winnipeg, MB
Junior Software Developer ML Team May 2018 – May 2019
• Achieved 15% increase in classification accuracy by scaling preprocessing pipeline to support 3X more features
• Increased training efficiency by leveraging transfer learning; Utilized a pretrained transformer network
• Attained >90% accuracy detecting actionable points in emails by compiling a balanced training set of 1M samples
• Reduced complexity & runtime by refactoring core data preprocessing logic with compact NumPy & Pandas routines
Invenia, Winnipeg, MB
Junior Software Developer Labs Team May – Sept. 2017
• Identified supply and bid curve patterns using unsupervised learning on hundreds of thousands of transactions
• Forecasted demand of nodes in an electricity network by using SQL and Julia to combine multiple data sources
• Presented insights to teams of researchers by creating a pipeline to scrape, analyze, and visualize millions of logs
University of Manitoba, Winnipeg, MB
Undergraduate Researcher NSERC Award May – Sept. 2016
• Made the concept of measure zero more intuitive by creating visualizations of recursively constructed sets
***Skills***
Languages: Proficient in Python; Experienced with SQL, Julia, Java; Familiar with JavaScript, C/C++, R
Data & Machine Learning: Keras, TensorFlow, PyTorch, NumPy, pandas, scikit-learn, Matplotlib, NLTK
Other: Git, Jupyter, Colab, Docker, Unix/Linux, AWS (EC2/Lambda/S3/SageMaker), Bash, OpenGL
***Projects***
Facial Recognition App – High School Mentorship Program
• Taught CS and ML concepts while creating a computer vision model to serve real-time in an Android app
• Compressed image classification model ~60% by using model quantization, pruning, and TensorFlow Lite
Takuzu Puzzle Solver
• Simplified variable space from O(n2) to O(n) by formulating an efficient problem representation
• Solved using AC-3 algorithm and depth first search of the constraint graph by reducing branching factor to binary
***Education***
University of Manitoba, Winnipeg, MB
B.Sc. Honours in Computer Science – Mathematics Sept. 2014 – June 2019
• Graduated above the 98th percentile of the class; Four-time President’s Scholar
• Bernard Noonan Memorial Prize for high standing in an honours program
• Philosophia Mathematica Prize in Applied Mathematics to the top 3rd year math student