Andreana M. Rosnik, Ph.D.

LinkedIn GitHub

San Francisco, CA, USA

Computational scientist, fast learner, creative problem-solver with deep mathematical background. Passionate about making an impact in climate, health, or social justice.

Experience

Freelance, clients include: Invicta Medical and Reflecting Equity (Remote)

Data Scientist

08/2023-present

  • Takes raw sensor data and converts it into tabular form

  • Develops anomaly detection algorithms that incorporate signal processing techniques

  • Standardizes survey data for diversity, equity, and inclusion (DEI) data analysis

  • Communicates findings via written reports, data visualization, and live code demonstrations

  • Continues education via Coursera classes in deep learning and Tensorflow development

  • Technologies: Python (pandas, numpy, seaborn, scikit-learn, jupyter), git, bash

Atomwise, Inc. (San Francisco, CA/Remote)

Cheminformatics scientist II

04/2021 – 05/2023

  • Curated datasets of millions of drug molecules for use in machine learning model training

  • Designed, trained, and validated machine learning (ML) models to improve molecular pose/binding mode prediction, a step in drug discovery modeling pipelines

  • Lead the five-person project to improve pose quality and prediction

  • Trained ML models, created visualizations, and designed chemically intuitive benchmarks for model explainability; presented this at the American Chemical Society Meeting in Spring 2022

  • Customized virtual high-throughput screening (vHTS) models for specific protein classes

  • Continued education via Coursera classes in deep learning and project management

  • Technologies: Python (pandas, numpy, jupyter, matplotlib, seaborn, sklearn), RDKit, MySQL, gitlab, bash, AWS S3, PyTorch

  • Domain knowledge: protein structure, organic chemistry, vHTS, molecular docking, AlphaFold2

Enel X (San Francisco, CA/Remote)

Software Engineer/Technical Lead, Distributed Energy Resources

09/2019 – 11/2020

  • Developed R Shiny web app for pre-screening solar + storage commercial and industrial projects that was used to assess up to 54 MWh total in new storage contracts in 2020

  • Orchestrated software releases and conducted code reviews for a five-person team in a fast-paced agile, test-driven development framework

  • Revised financial calculations for battery prices, service costs, and NPV to be < 2% of spreadsheet model results

  • Collaborated with product management to scope upcoming features

  • Technologies: R (shiny, data.table, testthat, renv, devtools), git, AWS CodeBuild, Docker, bash

  • Domain knowledge: demand response, energy tariffs, linear optimization

Enel X (Remote)

Optimization Engine Intern

06/2018 – 08/2018

  • Designed high and low precision optimization schemes in Java, saving up to 1-5% in total energy costs for simulated versions of existing commercial and industrial solar + storage projects

  • Created Python tools to visualize KPIs and identify irregularities in energy cost components

  • Technologies: Java, git, Python (numpy, matplotlib, scipy), bash, JIRA

University of California-Berkeley (Berkeley, CA)

Graduate Student Researcher

08/2014 – 08/2019

  • Solved non-linear equations and Monte Carlo simulations to predict organization in photosynthetic membrane stacks via a statistical mechanical modeling framework

  • Collaborated with five scientists in an international project to model photonics of higher-plant photosynthetic membranes

  • Received awards totaling ~$126,000 + tuition, inc. National Science Foundation Graduate Fellowship

  • Technologies: C++, Python (numpy, matplotlib, seaborn), bash, LaTeX

  • Domain knowledge: statistical mechanics, Monte Carlo simulations, enhanced sampling, lattice/Ising-based models

Education

Ph.D. Chemistry; Emphasis: Physical Chemistry (University of California, Berkeley; 08/2019)

B.S. Chemistry, Mathematics; Minor: Spanish (Hope College; 05/2013)

Portfolio

“Adventure Scrape”

06/2019

  • Scraped Wiki FANDOM Adventure Time transcripts and created tabular dataset from cleaned HTML data

  • Explored dataset via text mining

  • Designed character-level recurrent neural network to generate speech for a protagonist

  • Wrote a Medium post featured in Towards Data Science to communicate the technical findings to technical and non-technical readers

  • Technologies: Python (pandas, numpy, Beautiful Soup), PyTorch, Google Colab

Awards

  • 10 Under 10 Hope College Alumni Award (2019)

  • Outstanding Graduate Student Instructor Award (2017)

  • National Science Graduate Fellow (2013-2019)

  • Fulbright Research Fellow at Universitat de Barcelona (2013-2014)

Volunteering

Miscellaneous

  • Languages: English (native speaker), Spanish (professional proficiency), Catalan (limited proficiency)

  • Freelance fantasy illustrator featured in various group gallery shows in the San Francisco Bay Area (2015-present)