Andrew Bauer Email & Phone Number

Vienna, Va, Us

• Implement model serving pipelines using gRPC and pub/sub methodologies, alongside HTTP protocols, for robust and scalable deployment of machine learning models.• Leverage open-source LLMs for XAI and re-architect and post-train/fine-tune them (e.g. LoRA) on our datasets for specific tasks.• Manage, mentor, and collaborate with other team members in integrating these technologies for optimal production performance.

Vienna, Va, Us

• Build an anti-fake news grader (news websites and podcasts) that upweights transparent and reliable sources of information using NLP:https://www.seekr.com & https://www.seekr.com/seekr-align.• Company's previous name -- NTENT: https://ntent.com

Bozeman, Montana, Us

• Developed machine learning models for anti-money laundering, and preprocessed/analyzed data, using Python, Java, and Pyspark.• Developed supervised classifiers that achieved 90% recall at 1% FPR at detecting money laundering through corporate bank accounts in the Asia-Pacific region (helped delivered first ant-money laundering product at FICO).• Helped develop unsupervised clustering algorithms using NLP that achieved 90% recall at 4.7% FPR at detecting corporate money laundering, and that discovered new money laundering patterns.• Worked with our clients to ensure proper data mapping and delivery through assessments of data quality (mapping between client's data and our schema, detection of missing/invalid values, etc.).• Wrote Java production code to support different features of our product, e.g. adding functionality to provide quantile estimates for different variables (i.e. providing quantitative measures of anomalous/outlier behavior).• Solely launched a Jive forum for the analytics department to promote interaction during COVID-19 remote work.

San Mateo, Ca, Us

• Applied machine learning to credit/debit card and wholesale banking fraud detection, and mined data for business insights using Python (scikit-learn, pandas, numpy, scipy), Hive and Impala, Spark, and Pulse (Feedzai's proprietary ML platform).• Lead data scientist on Citibank (client) project: developed classifiers that achieved 66.8% recall and 99.7% money recall (at 0.2% FPR) in detecting wholesale banking fraud ($1 trillion USD dataset); worked cross-functionally and translate findings for business audiences.• Developed classifiers for First Data (client), a major acquiring bank that processes 50% of all US debit and credit card transactions.

San Francisco, Ca, Us

• Consulted for a fintech startup to improve fraud detection on its mobile payments app.• Developed an ensemble classifier using deep learning (autoencoder) and other classifiers (random forest, gradient boosted trees) for fraud detection using Python (keras, TensorFlow).• Reduced incidence of chargeback fraud to 0.2% of total transactions, a 7% increase in the company's previous classification performance.

Toronto, Ontario, Ca

• Designed experiments and analyzed behavioral data collected online from Amazon Mechanical Turk.• Implemented mixed effects regression and ANOVA models using Python, SQL, and R.• Obtained competitive Brain Canada grant based on merits of proposed research ($100,000 CAD).• Managed a team of undergraduate researchers and taught data science in Python to lab members.

Pittsburgh, Pa, Us

• Designed cognitive neuroscience studies and analyzed large neuroimaging datasets.• Applied machine learning to brain activation data using supervised (SVM, logistic regression) and unsupervised (factor analysis, k-means) methods in MATLAB and R to decode a person’s thoughts.• Published 4 first-author articles in top-ranking journals; some of this research has been featured in media outlets such as Wired and KurzweilAI.

Pittsburgh, Pa, Us

• Developed and taught an upper-level undergraduate lab course on applying machine learning and statistical methods to neuroimaging data for 3 semesters (~20 students each semester).

Bloomington, Indiana, Us

• Recruited and ran participants in behavioral and fMRI experiments.• Performed machine learning classification of brain states in MATLAB to reveal what a person is imagining and feeling.

Doctor Of Philosophy (Ph.D.), Cognitive Neuroscience

Master Of Science (M.S.), Cognitive Neuroscience

Bachelor Of Science (B.S.), Neuroscience, Psychology (Double)