András Novoszáth

Data Scientist, Freelancer, multiple clients (09/2018 – present): - Data Analytics: informed technical & business decisions. - Machine Learning: predicted variables and evaluated model performance. - Engineering: built data processing pipelines and runnable packages. Notable projects: User recognition algorithm - Built model to recognize individual user sessions responsible for 0.5% of the whole dataset. Machine learning benchmarking - Built machine learning modeling pipeline to benchmark feature engineering with multiple configurations, techniques, and datasets. HVAC time series data analysis - Built HVAC consumption predictor and mold probability estimator based on humidity and temperature data. Glucose level forecasting - Improved the original prediction algorithm to surpass clinical-level prediction criteria and meet the prediction power of state-of-the-art non-mobile diagnosis tools. Medical IoT data analysis and reporting - Built a customizable reporting pipeline from clinical trials data updating decision-makers about week-to-week results. - Insights helped the project to secure a 8-figure funding round.

Worked in a fast-growing team developing a wealth management platform from prototyping to launch. - Improved development processes by building and maintaining +500 pages of modular, customer-facing documentation (dashboards, KPIs, used metrics). - Communicated value to prospects by writing RFIs, RFPs, and Q&A for proposals. - Upscaled documentation by coordinating 2-4 technical writers and introducing single-source tools. - Uplifted brand recognition by writing three white papers on wealth management, mobile banking, and financial innovation.