Maximilian Klein Email & Phone Number

Albany, CA, US

Spearheaded data standardization and re-architecture to merge multiple ML workstreams (Civic, Climate, Health) for model-performance gains, and developer efficiency. Lead implementation of new Machine Learning pipelines, scaling sampling, labelling, and measurement facilitating the training of 10s of models on millions of human-coded labels.

Working with the Social Impact team, I support the Civic Graph and Civic Classifier initiatives.

Developed and deployed core modules, integrations, and web-apps enabling our citizen science A/B testing service for Wikipedia and Twitter users. Owned and assembled datasets performing power-analysis, hypothesis testing, and data exploration for 15+ field experiments. Established Data Science practices, workflows, and technologies for collaboration.

San Francisco, CA, US

Spearheaded project using Machine Learning to improve user retention. Executed concept end-to-end: trained and modelled AI automatically detecting goodfaith users, then deployed a realtime bot that connects them with mentors. Validated the AI in live experiment on English Wikipedia, boosting user-retention by 20%.

Engineered features for machine learning models in credit-risk tasks, deriving complex attributes from external and internal APIs. Hardened development code through testing to live production environments in fraud detection.

Livermore, CA, US

In the global security team, implemented a tool for identifying rare aircraft in satellite imagery by training convolutional neural networks on CAD renderings, solving an “example-less” computer vision problem.

Minneapolis, Minnesota, US

Member of GroupLens research group under advisor Prof. Haiyi Zhu. Focuses on peer-production and the sharing economy using a human-computer interaction lens and data science techniques.

In partnership with Tulsa Public Schools, created an Early Warning Indicator of performance for 3rd Grade students. Utilizing data science and machine learning, we developed a predictive model catching 95% of at-risk students.

San Francisco, CA, US

Won $22,500 grant to create Wikidata Human Gender Indicators, a gender bias live tracker over all Wikipedias. Headed a team of 3 developers in engineering the first analytics dashboard for oversight of Wikidata. Statistics generated by the project have been cited in the BBC, Bloomberg, and on the cover of Wikimedia annual report.

Developed Web App (Python) that watches changes to Wikipedias citations in real time, then converts and uploads qualifying Open Access Papers to Wikisource for deeplinking, redundant citability. Meanwhile analysing usage trends in citations of Open Access Research in Wikipedia.http://notconfusing.com/the-topmost-cited-dois-on-wikipedia/

Dublin, OH, US

Bridging the gap between library data and Wikipedia. Racked up over 2 million Wikipedia edits with custom-written bot to improve articles using Library data.Blogging on the subject: http://hangingtogether.org/?cat=51

Oakland, CA, US

Taking Open Culture so seriously that I became the one member of the hackerspace that could be held legally liable. Backstoping all financial responsibility of the hackerspace, while performing some data analysis on donations at the same time.In other capacities, have been teaching prerequisite-free math and 3D priinting.

Doctor Of Philosophy (Phd), Computer Science

Master Of Science - Ms, Computer Science

A.B., Mathematics