Machine Learning Data Science

Duncan A. Clark, PhD Milwaukee, WI (USA) 937-***-**** **********@*****.*** LinkedIn GitHub

Qualifications Summary

Machine learning specialist with PhD in Mathematics and strong foundation in data science and statistical modeling. Proven ability to communicate complex concepts to diverse audiences through 10+ years of teaching and public speaking. Experienced in deep learning and building predictive models using Python. Skills

- Python; SQL; VS Code; git; MS Word, Excel, PowerPoint; Windows, Linux, MacOS; Jupyter Notebook

- Hard skills: Data cleaning, probability and statistics, supervised/unsupervised learning, predictive modeling, neural network design and traning, ensembling, boosting, clustering algorithms, natural language processing

- Python packages: pytorch, numpy, scikit-learn, transformers, xgboost, pandas, matplotlib, seaborn, gradio

- Soft skills: Self-motivated learner with sharp problem-solving skills, excellent written and oral communication Professional Experience

Milwaukee School of Engineering Milwaukee, WI

Instructor Aug 2021 – Present

- Designed and implemented curricula for nine undergraduate courses per year (20-25 students each). Planned interactive modules and created group exercises, assessments, and custom digital textbooks

- Organized and administered large regional mathematics competition for 100+ high school students, designed problem and solution sets, managed event logistics and volunteer coordination The Erdos Institute

Teaching Assistant for Data Science Bootcamp May 2025 – Present

- Led small group problem solving sessions introducing cohort of 100+ PhD students and recent graduates to data science techniques and methods through guided Jupyter notebooks Ohio State University Columbus, OH

Teaching Associate & Research Associate Aug 2015 – May 2021

- Acted as lead instructor and teaching assistant for undergraduate math courses, designed lectures and assessments and provided student support, and earned a university-wide award for excellence in teaching

- Three-time recipient of RTG National Science Foundation Research Training Group, valued over $40,000 Projects

FSP Finder GitHub Hugging Face Spaces video May 2025 - Present Description: FSP Finder is an AI-powered tool for detecting and automatically censoring explicit content in music lyrics. A convenient web interface allows the user to upload one or more tracks and receive (1) a full transcript of each song along with (2) a high-quality edited copy with the vocals muted at times detected to have explicit content

- Gained valuable experience in fine-tuning OpenAI Whisper on large datasets and front-end design with Gradio

- This tool is designed to be implemented for local radio station WMSE-Milwaukee to determine if tracks are suitable for airplay and streamline the process of editing explicit tracks Predicting ratings from review sentiment GitHub Jan – April 2025

- Trained recurrent neural network to predict review ratings and sentiment from review text. Extracted over 20 million restaurant ratings and reviews from US Google review data as training dataset.

- Achieved over 98% model accuracy on sentiment prediction, and 94% accuracy on star-rating predicting, matching or exceeding transformer-based models such as BERT or RoBERTa with 10x less training time Education and certificates

Ohio State University PhD in Mathematics, specialty in Algebraic Topology and Homotopy Theory May 2021 Ohio State University BS (w/ honors) in Mathematics May 2015 The Erdos Institute Data Science Bootcamp (received distinction) Jan – Apr 2025 The Erdos Institute Deep Learning Bootcamp May -- Aug 2025 Publications

On the Goodwillie derivatives of the identity in structure ring spectra, Tbilisi Math. Journal (2020) available here

- Solved novel problem in algebraic topology, published in special issue of a peer-reviewed mathematics journal

- Communicated results at 12+ invited conference talks and research seminars

Contact this candidate