PETER RASMUSSEN

Data Scientist

pgr-me@protonmail.com | linkedin.com/in/pgrme | github.com/pgr-me

PROFILE

I’m a data scientist whose multi-disciplinary background equips me with a broad set of skills to solve complex, unstructured problems.

SKILLS

Programming

Python | SQL | Bash | Git | AWS

Machine Learning

Regression | Classification | Clustering | NLP | Time Series

Big Data

Hadoop | Hive | Impala | HQL

Python Libraries

Geospatial Tools

QGIS | Carto | PostGIS | Mapbox | GDAL

EXPERIENCE

Accenture Federal

2023-Now Analytics Manager Washington, DC

IARPA Walkthrough Rendering from Images of Varying Altitude (2023-2024) Leading effort to develop method to register disparate satellite, aerial, and ground imagery.

2019-2023 Analytics Associate Manager Washington, DC

IARPA Space-Based Machine Automated Recognition Technique (2021-2023) Leadership of several modules in critical path played a key role in successfully reaching second phase of program. This success has contributed to follow-on research wins and serves as a credential in support of broader national security portfolio growth. Led two university teams & junior Accenture staff, integrating teams’ algorithms into pipeline under tight deadlines. For second client independent evaluation, made critical contribution by implementing innovative optimization routine. For third independent client evaluation, achieved step change improvement in scores by integrating outputs of both models.

USDA COVID Analytics Team (2020-2021) Oversaw team’s data science and engineering operations. Developed ETL job to load Census data into enterprise data warehouse. Trained regression model using loan data to understand drivers of loan delinquencies.

Census Self-Response Quality Assurance (2020) Trained supervised learning model to detect bot activity, synthesized analytical outputs into presentation, and presented actionable findings to customer. Designed and developed Python machine learning pipeline to ETL external and internal datasets that fed into unsupervised learning model. Created ETL pipeline in Python to move tens of millions of records from Hadoop to Oracle to facilitate subsequent analysis in Oracle.

Commute Optimization Project (2019) Designed and developed software that recommends - using a bipartite matching algorithm - an optimal set of intra-company role swaps that minimize commute times. Organized data processing into discrete tasks using a Python-based workflow automation tool: acquisition, cleaning, routing, post-routing processing, recommendation, and reporting.

Gallup

2019 Lead Data Scientist Washington, DC

Automation of Survey Data Processing Designed data processing pipeline for global, multi-million dollar polling project to automate country-level survey data cleaning processes that used to be carried out manually. Impact: Processing time reduced from up to three months to less than three hours per survey.

Hierarchical Personality Clustering Trained ensemble of dimensionality reduction and hierarchical Gaussian mixture models to cluster personality assessment data into hierarchically organized personality types.

Automating Scientific Knowledge Extraction As part of an AI project for a defense research entity to create an ensemble of models to generate hypotheses for a domain of knowledge, I 1) built a research article scraping tool to retrieve and process open source papers and 2) developed a script to parse Markov Logic Network data and store it in a graph database.

Hierarchical Engagement Visualization Tool Created JavaScript tool that uses D3 to visualize employee engagement hierarchically. That is, manager-to-manager engagement relationships from CEO down to the lowest level manager. Decision-makers can use tool to clearly see how the engagement of one manager affects their subordinate managers’ engagement.

Understanding Group Biases As part of an AI solution for a defense research entity to understand group biases, designed and coded a web scraping tool to retrieve, organize, and assimilate reader comments data from popular press articles across the political spectrum.

2017-2019 Application Developer Washington, DC

Price Optimization Model Influenced senior leadership to drastically change prices of signature products with potential multi-million dollar YoY impact on revenue based on results from integrated time series and econometric model.

Job Title Classifier Built logistic regression job title classifier using NLP techniques to generate standardized set of job types, which we used as data inputs to score marketing leads.

Analysis of Global Coaching Market To inform senior leadership’s strategy regarding coaches, led the quantitative analysis of the global coaching market. Identified key market segments and quantified served and addressable portions of each. Organized findings into presentation and delivered my portion of the work to key stakeholders.

Strengths Coach Bot Key member of initiative to prototype an Amazon Alexa-powered coach chat bot. My contributions included 1) connecting each piece of the application together so that it functioned as required; 2) debugging code; 3) meeting with senior leadership to discuss strategic options to continue to develop the prototype; 4) meeting with prospective client at its annual leadership retreat to discuss how the prototype could prove beneficial; and 5) demonstrating utility of prototype to subset of retreat attendees, who by and large liked the prototype.

Ongoing Ecommerce Reporting Created data acquisition, cleaning, and analysis jobs in Python to generate reports for key stakeholders on various aspects of firm’s ecommerce business.

MITRE

2016-2017 Senior Data Scientist McLean, VA

Disability Claims Simulation Conceptualized and developed a discrete event simulation in Python to model disability claims processing for US Department of Veterans Affairs.

Metis

2016 Data Scientist New York, NY

Revealing Economic Growth Using Night Light Imagery Used satellite night light imagery data to interpolate gaps in GDP data in countries for which this information was unavailable.

Marijuana through the Lens of the New York Times Explored how the New York Time’s view on marijuana has evolved using NLP, dimensionality reduction, and clustering techniques to semantically group articles into epochs. Identified key words in each cluster based on TF-IDF results and compared cluster words to trends in public opinion and historical events. Blog post featured in Data Science Weekly.

Visualization of Water Affordability Used univariate (Jenks) and multivariate (K-Means) clustering to understand stratification of water pricing and consumption in the US using Census data.

Understanding Loan Activity among Microlenders Built an explanatory linear regression model for an NYC microlending company to identify types of people who are most active microlenders.

HIP Consult

2013-2016 Analyst Washington, DC

Fiber Optic Infrastructure Web Mapping Led product development of SaaS infrastructure mapping prototype. Managed GIS team tasked with assimilating and mapping approximately 350,000 km of fiber optic infrastructure for 42 countries across Africa.

Digital Finance Market Analyses Led quantitative analysis to develop value propositions for a multinational company seeking to enter digital finance markets in several African Markets. Financially modeled market potential and expected cash flows of Tanzanian mobile payments startup. Authored market-sizing and forecasting section of a commercial report on the characteristics, forecasts, and trends in MFC for several African markets.

South Africa Fiber Backbone Infrastructure Project Analyzed key, critical-path operational issues and tracked progress to facilitate timely project completion. Developed financial models to assess cost to client and conducted quantitative analysis to facilitate resolution of key strategic issues.

Chase Construction Limited

2011-2012 Project Manager Nairobi, Kenya

Led research team in analyzing investments in renewable energy and agricultural technologies, from conducting fieldwork in Nairobi and Mogadishu to market research, engineering design, and financial evaluation of potential projects. Co-organized first TEDxMogadishu, a series of talks given by Somalis on the “rebirth” of Mogadishu, including the resurgence of the private sector.

EWB

2010-2011 Program Manager Amherst, MA

Managed Engineeers Without Borders (EWB) UMass Kenya program, which focused on improving the water quality in western Kenya for a rural community of approximately 10,000.

UMass Amherst

2010-2011 Research Assistant Amherst, MA

Modeled long-term economic impacts of technical change in carbon capture given various greenhouse gas emissions policies. Conducted sensitivity analyses of key performance drivers of carbon capture costs to simulate the impact of technical change on cost.

2009 Teaching Assistant Amherst, MA

ERM

2006-2009 Staff Engineer Metairie, LA

Managed environmental consulting projects across multiple sectors. Successfully liaised between regulatory organizations and clients. Routinely developed reports, presentations, and proposals.

University of Iowa

2002-2005 Research Assistant Iowa City, IA

Conducted research on microemulsions and their applications in the the development of environmentally-friendly photopolymers.

EDUCATION

MS

2024 Johns Hopkins University Artificial Intelligence

Graduated with honors

Certification

2016 Metis Data Science Bootcamp

MSIEOR

2012 UMass Amherst Industrial Engineering & Operations Research

Thesis: The Economic Impacts of Technical Change in Carbon Capture

BSE

2006 University of Iowa Chemical Engineering

Graduated with honors, minor in chemistry

PUBLICATIONS

Xiaojing Tang, Madison G. Barrett, Kangjoon Cho, Kelsee H. Bratley, Katelyn Tarrio, Yingtong Zhang, Hanfeng Gu, Peter Rasmussen, Marc Bosch, Curtis E. Woodcock (2024). Broad-area-search of new construction using time series analysis of Landsat and Sentinel-2 data. Science of Remote Sensing.

Rasmussen, P., Abrahamson, J., Tang, X., Smith, O., Gray, J., Woodcock, C., & Bosch, M. (2023, July). Assessment of Performance of Tree-Based Algorithms to Reduce Errors of Omisssion and Commission in Change Detection. In IGARSS 2023-2023 IEEE International Geoscience and Remote Sensing Symposium (pp. 6676-6679). IEEE.

Baker, E., Nemet, G.F., and Rasmussen, P. Modeling the Costs of Carbon Capture. Handbook of CO₂ in Power Systems. Ed. Zheng, Q.P. et al. Springer, April 2012.

Jain, K., Rasmussen, P., Rethwisch, D.G., and Scranton, A.B. Enhanced Epoxidation of Soybean Oil through Microemulsion Technique. Polymer Preprints, 45(2), 579-580, 2004.