Machine Learning Data Scientist
Resume:
Logan MacFarland
Beaverton, OR 97006
*****.**********@*****.***
EXPERIENCE
Riise.ai, Remote - Full Stack Developer
Aug 2024 - Jan 2025
● Lead the development of a computer vision platform using Python and Pytorch that analyzes satellite imagery and public datasets to estimate the potential value of building upgrades and their impact on energy use and emissions using just an address.
● Implemented a chatbot feature that allows users to ask questions in plain language and receive clear answers based on model-generated insights. The chatbot can also generate reports and recommend next steps based on the data. Chatbot extracted data directly from the platform.
● Oversaw the entire technology stack, including front-end and back-end development, data processing, and machine learning models, while ensuring the platform runs efficiently and securely.
● Built a website using React that displays analyzed buildings on an interactive map and presents key statistics such as return on investment, cost estimates, and projected emissions reductions.
● Worked closely with investors and clients to align the platform with market needs and sustainability goals.
Outlier, Remote Contract - Data Science Consultant Feb 2024 - Aug 2024, Jan 2025 - Current
● Collaborated on projects that involved generating executable Python and SQL queries in response to non-code, plain English requests submitted to Outlier’s platform.
● Assisted in refining AI models to accurately interpret and respond to human code requests, ensuring the returned code met the desired functionality.
● Provided clear explanations alongside the code outputs, helping users understand the logic and execution of the generated solutions.
First Bank of the Lake, Remote - Data Scientist
Mar 2023 - July 2023
● Developed a custom machine learning model using Python’s XGBoost package to estimate the average lifespan of loans, optimizing purchasing decisions to maximize profit within the loan purchasing program,
● Engineered features by leveraging various public datasets, ensuring the model handled inconsistent input and missing data to provide real-time recommendations.
● Integrated data from SQL databases, XML files, and web scraping from third-party sources to support model predictions, with the solution deployed via AWS SageMaker notebooks.
● Collaborated closely with market experts to identify key purchasing points, using model outputs to determine optimal bid amounts for loans in the open market. Iteratively adjusted bidding strategies based on historical market feedback.
Credibility Capital, Remote - Data Scientist / Data Engineer July 2022 - Feb 2023
● Created a centralized data lake to allow machine learning models to simultaneously reference multiple data sources, using AWS Lambda, AWS AppFlow, and AWS Glue for automated upserts.
● Optimized AWS infrastructure, reducing operational costs by minimizing resource uptime and improving resource allocation.
● Built and maintained data pipelines between AWS and Google Cloud using automated SQL queries, ensuring efficient data flow and accessibility.
● Led database design, indexing, and maintenance for the SQL-based data lake, creating scripts to streamline data access for performant model generation in AWS.
● Took on dual responsibilities, managing Data Engineer tasks while progressively building towards Data Science tasks, supporting the company’s evolving infrastructure and machine learning needs.
Goldstar Events, Portland, OR - Data Scientist
Mar 2019 - Feb 2022
● Developed a novel machine learning model to represent user preferences, implementing a newly published variant of factorization machines that leveraged item trends and regional data. Involved end-to-end infrastructure development, feature selection, memory optimizations, and data pipeline creation. Model was optimized to only require minimal user data and dynamically represent user preferences as a user interacts more with the platform.
● Deployed machine learning solutions directly to AWS using S3, Lambda, EC2, SageMaker, and Athena, ensuring scalable and efficient production systems.
● Built an item similarity model using gradient boosting, incorporating NLP data, user behavior patterns, and categorical features. Model greatly increased item to item browsing on platform.
● Shifted focus during the pandemic to address urgent business needs, including data migration, market analytics, and data science consulting for external clients.
● Led a major data integration effort between Goldstar and its acquiring company, TodayTix, utilizing text processing techniques to match and merge duplicate event listings across both platforms.
Nike, Hillsboro, OR - Data Analyst
Jan 2019 - Mar 2019
● Developed an algorithm using Python and R to fuzzy match addresses across multiple databases, accounting for data inaccuracies and inconsistencies in large datasets.
● Designed and implemented data pipelines for querying, cleaning, and assigning matches within big data environments, ensuring data integrity and scalability.
● Created a scoring system to evaluate the strength of address matches, using hand-labeled data to tune the algorithm for optimal performance.
● Automated the matching process, with the algorithm either assigning correct matches or flagging ambiguous cases for manual review, significantly improving accuracy and reducing manual effort. First Tech Federal Credit Union, Hillsboro, OR - Data Analyst Nov 2017 - Dec 2018
● Extracted, cleaned, and transformed vendor-supplied XML data describing member assets into a tabular format, processing text descriptions to identify company names and asset types for further analysis.
● Modeled deposit balances using Kaplan-Meier survival analysis, leveraging a survivorship model with undefined endpoints to maximize sample size and improve accuracy in predicting member behavior.
● Enhanced reporting processes inherited from a predecessor, automating and optimizing reports while collaborating directly with directors and project managers to align with business objectives.
● Conducted various big data ad hoc analyses using R, Excel, and Tableau, drawing from multiple data sources including internal SQL databases (Microsoft and Oracle), public datasets, data warehouse tables, Salesforce, and vendor databases. Visualized data using plotly (similar to Seaborn)
Selectron Technologies, Portland, OR - Operations Developer April 2016 - May 2017
● Automated data retrieval from Outlook, SQL databases, and enterprise software APIs, making reports more accessible and efficient.
● Analyzed and visualized company-wide data for quarterly reviews using R and PowerShell.
● Built a statistical model to forecast system call volume, improving resource planning.
● Designed and maintained a centralized SQL database, automating data imports and ensuring clear documentation.
● Optimized legacy SQL procedures, enhancing report accuracy and performance. EDUCATION
Oregon State University - Bachelor’s in Microbiology
Contact this candidate