Henry Daniels-Koch Email and Phone Number

• Trained and deployed NLP sentence classification and named entity recognition models into Amplion's data pipeline using transformer based BERT techniques on medical publications and NIH grants. This model output is the most widely used saved filter across all Amplion customers.• Modeled synonym “quality” with a decision tree based model to detect and prevent problematic synonyms from being automatically imported into Amplion’s synonym ontology, significantly reducing manual curation• Predicted success of preclinical drugs with a model using features generated from NLP classifiers• Utilized machine learning models including random forest, regression, SVMs, LSTMs, and neural networks for a variety of trial projects.• Took on a software engineering role in the backend (Python, Go) to deploy new models, extract and clean unstructured data, and execute on key features• Led data heavy features in a product manager role on design, requirements, and story creation

• Developed dashboards with Looker visualization software with BigQuery as key customer deliverables as the only developer in this product line.• Improved data quality through diagnostic dashboards in jupyterlab and led data curation initiatives• Explored, deployed, and maintained fuzzy matching algorithms to boost our synonym matching accuracy for organization names.• Created custom data deliverables with BiqQuery

• Analyzed cost and utilization of health services for internal and external reports using SQL, SAS, and Tableau• Designed QA test scripts to ensure databases are following quality measure specification requirements• Created reports on primary care spending from claims, capitation, provider incentives, and risk withhold

• Represented 350Deschutes at the Low to Moderate Income (LMI) Solar Working Group monthly meetings in Portland, which aim to bring solar power and solar jobs to LMI populations• Presented LMI Solar meeting educational content at 350Deschutes' Renewable Energy Committee weekly meetings• Campaigned for Oregon's Clean Energy Jobs Bill at businesses in Bend

• Numerically simulated early universe energy fluctuations using finite differencing methods on a non-linear partial differential equation in C++ with spatial and temporal discretization• Linearized the equation and solved with predictor-corrector method in 1D and 3D in spherical polar coordinates• Determined initial conditions using both a shooting method and a relaxation method of an ODE

• Conducted research with NSF grant titled "Numerical Simulations of Black Holes"• Parallelized numerical relativity code in C++ using OpenMP in order to computationally simulate black holes, neutron stars and gravitational radiation• Worked closely with another student and communicated ideas clearly and efficiently each day to PI

• Graded physics homework on a weekly basis

• Reduced optical and infrared data from the Keck Observatory telescope using IDL code with IRAF software• Produced an emission spectrum using IRAF for 8 quasars• Analyzed and used academic papers related to Dust-Reddened Quasars• Modified C++ code to produce Gaussian fits that approximated fluxes of emission peaks• Plotted BPT diagrams to determine whether objects were quasars• Presented research at KNAC Consortium Symposium at Williams College in 20 min talk

• Wrote code in Matlab that analyzed ocean currents, wind speed, shipping traffic density, and piracy zones in order to plot shipping routes that minimized clients' fuel costs.• Created a GUI user interface that found relevant charts and data for routes created• Translated data into Mercator projection coordinates