Student Employment & Internships

Corteva Agriscience™ has an exciting internship opportunity at UIUC campus during Summer 2020. The successful candidate will work with leading research and development scientists to develop and apply state of the art statistical and machine learning methods in the area of bioengineering and bioprocessing. Project: Impact of data partition on fermentation optimization Required Qualifications Currently enrolled in a graduate program in quantitative sciences such as Statistics, Computer Science, Bioinformatics or a related field. Experience with statistical inference and prediction modeling. Knowledge of descriptive statistics and experimental data analysis, particularly analysis of variance, mixed models and non-linear regression. Programming experience in R or Python, preferably in high-power computing environment. Effective written and verbal communication skills. Candidate should demonstrate ability to successfully collaborate with Scientists. Attention to detail is a critical skill for the internship. Candidate should be able to develop accurate software code and written reports. Critical thinking and strong problem-solving skills are a plus. This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks, Lisa.DunganBrooks@corteva.com. Please reference “Impact of data partition on fermentation optimization Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Deep Learning Accelerated Manifold Learning for Reactive On-line Environment Classification Project Goal A user interface and machine learning method enabling real-time, deep-learning-accelerated manifold exploration and learning to compress GxExM datasets into environment classes. Background and Motivation Genetics by environment by management datasets are being collected to develop environment and mega environment classifications for breeding program optimization. These datasets consist of highly multivariate and colinear variables. Optimal and actionable segmentations of environment are needed to provide breeding teams with environment targets for selections and to compare the rank genetic materials in muti-location environmental trials. These activities will occur within and across breeding evaluation zones and continental business structures. Furthermore, the genetic material available will change yearly and the problem differs significantly by crop. The dynamic factors of spatiotemporal environment and genetic variability will require a dynamic capability to efficiently compress and solve optimizations across the interacting objectives of environment class and product concept. Static envirotyping approaches will be effective only for static genotype x environment realities. Project Description The candidate will use GxExM datasets for NA corn breeding locations since 2012 and apply deep learning approximations of multiple manifold learning techniques to compress the dataset into meaningful latent spaces. The candidate will develop a live exploratory data analysis tool powered by these methods to enable rapid visualization and conceptualization of envirotypes and their match to product categories. This will be used to drive semi-supervised classification of environments. A further goal will be to explore the stochastic representation of environment possibilities and techniques to match the probabilistic outcomes to geospatial product placement. Deliverable(s) and Impact: An exploratory data analysis UI backed by a deep learning accelerated architecture that will enable advanced breeding decision makers to explore environment data spaces. The tool and analysis will be used by breeders and discovery breeders to evaluate target populations of environment for EZ and mega-environment applications. The tool could enable exploratory TPE targeting, optimization of GxExM, and development of environmental classifications in an on-line and dynamic system. Desired Skill Set Strong knowledge in deep learning, statistics, manifold learning, or machine learning Python, R, docker, linux Dash, Shiny, or other hardware accelerated multivariate data visualization systems Geospatial data manipulation and analysis a plus This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks, Lisa.DunganBrooks@corteva.com. Please reference “Deep Learning Accelerated Manifold Learning for Reactive On-line Environment Classification Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Corteva Agriscience™ seeks students for full-time internships within our global Data Science and Informatics group working on rapid imaging solution development for digital agriculture at our University of Illinois at Urbana-Champaign offices. We are aggressively building Big Data and Predictive Analytics capabilities to deliver improved services to our customers. Our interns will work on key problems related to delivery of imaging technology applications for internal phenotyping assays. The intern should be familiar with software development tools and practices on Linux/UNIX environments. Intern will work with imaging scientists, machine learning engineers and clients to provide customer value. What you’ll do: Design and implement web-enabled customer-facing frontend and backend software development for clients to consume imaging-related assay outputs. You will utilize front-end web development tools and frameworks such as React, JavaScript, Bootstrap, VUE.js, etc. to create easy-to-use user experiences. Develop appropriate backend infrastructure with tools and languages such as Python, SQLAlchemy, Pandas, OpenCV, SciKit-image, NodeJS. Create interfaces against production data repositories to consume imagery and related metadata. Enable scientific software development by following best practices and employing basic project scaffolding approaches (semantic versioning, security, infrastructure-as-code, etc.). Meet and work closely with clients to get requirements and feedback on solutions. Education and experience: Currently enrolled in an undergraduate education program pursuing (or planning to pursue) a B.S. in Computer Science, Software Engineering or related scientific discipline. Experience in C/C++ and Python programming with the ability to quickly create prototype solutions on Unix / Linux platforms. Development with front-end and back-end web technologies e.g. service-oriented architectures, React, AngularJS, Flask, FastAPI, MongoDB, REST APIs, JSON, HTML, JavaScript, CSS, NPM, grunt, gulp, basic project scaffolding. Strong knowledge of Linux with the ability to be able to build applications via the command line, debug issues, update software on Linux servers & perform system administration tasks. Proficiency with bash scripting and CLI tools (awk, sed, find, xargs, wc, head). Experience with Docker, application containerization, deployment and orchestration. Interest in learning new technologies, programming techniques, languages, and operating systems is a must. Excellent interpersonal skills and a can-do attitude with the ability to thrive in a fast-paced dynamic environment. Excellent analytical and problem-solving skills with the ability to work as part of a global team and, at times, independently while appropriately prioritizing tasks. Strong verbal and written communication skills in English are required. This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) Rapid Imaging Solutions Developer Project Goal Working closely with project stakeholders, enable client-facing web interfaces and related solutions to display and consume image data for imaging related assays. Create easily configurable and user-friendly interfaces with code reusability in mind so that solutions can be extended to other projects with minimal effort. Project Description The intern will be doing last-mile development. This includes UI mockups and development, integration or development effort required for value generation for the client. Deliverable(s) and Impact: What will be delivered at the end of the project: Last-mile development solutions. The interface, integration and development effort required for value generation for the client. Interfaces to consume imagery data based on common infrastructure. What is the impact that the deliverables are going to create: Enable digital transformation via enablement of AI/Computer Vision solutions to end clients. Digital Transformation. About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks, Lisa.DunganBrooks@corteva.com. Please reference “Rapid Imaging Solutions Developer Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Training and evaluation of multi-region remote sensing yield prediction models with additional remote sensing/large scale features Project Goal To determine the impact of implementing a multi-region yield prediction strategy on our ability to predict yield across varied environments. To determine the impact of adding more satellite features into the current yield prediction model Background and Motivation We currently utilize a single yield prediction model for predictions across the continental United States which has been trained on > 1 million corn, and >600K soybean fields. We have shown that we have reached a point in which adding addition training data will not significantly increase the accuracy of our models. The goal of this project will be to try several approaches to determine if we can improve the model’s accuracy by 1. creating new models based on state, CRM, evaluation zone, etc., or 2. adding additional features, satellite measured soil moisture, soil attributes, etc. Project Description Project Outline: Identify the zones to be tested and features to be added Extract the additional features and group data into the determined zones to create new training data sets Train/evaluate machine learning model Evaluate the performance of the model compared to the current approach Assemble and communicate the findings to the group This project will require knowledge of geospatial data structures, data manipulation tools (python, pandas, etc), remote sensing basics, machine learning, and data visualization. Deliverable(s) and Impact: What will be delivered at the end of the project: Clear understanding of the impact of a multi-zonal yield prediction strategy and additional features on accuracy. What is the impact that the deliverables are going to create: We hope to increase the accuracy of our remote sensing yield prediction models which can then be folded into current and future products which depend on this ability Desired Skill Set Knowledge of geospatial data structures, data manipulation tools (python, pandas, etc.), remote sensing basics, machine learning, and data visualization Machine learning experience does not have to be extensive, but candidate should be familiar with concepts Candidate should be familiar with: Python/R, pandas, geopandas, xgboost, multivariate linear regression, geospatial data visualization, scientific plotting, basic statistics This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Training and evaluation of multi-region remote sensing yield prediction models with additional remote sensing/large scale features Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Insect Resistance Modeling Project Goal Continue ongoing work in insect resistance, document results internally, and finalize publication for insect resistance. Background and Motivation The onset of insect resistance to insecticide products or insecticidal traits often renders the product concept useless when it occurs. It takes years to build up resistance that often comes from repeated and overuse of the pesticide or trait. The onset of insect resistance must be numerical since one is dealing with potential future behavior, along with estimates of future market share (for new products), conventional pesticide use, and use of Bt crops. Insects are known to develop resistance to conventional pesticides, and often various management practices are used to delay resistance onset. Modeling the impact of various management practices on resistance development provides the registrant details on how to control the pest and manage the onset of resistance such that the product concept can be fully realized. Ideally, a product concept should last at least 10 years. Once resistance is fully developed, the pesticide product concept is ineffective and over. Thus, understanding ways which can delay the onset of resistance helps when making recommendations for the company that maximizes profit and extends the product lifespan. Project Description This project uses modeling approaches (Python) to generate tools that describe insect resistance predictions and various management practices/scenarios that can be used to influence resistance behavior. Deliverable(s) and Impact Much of this work has been generated but now requires summarization and extensions. The intern will take and expand existing technology and document via internal and external publication. This work and generation of these tools propels Corteva agriscience to the forefront of industrial knowledge in the area of resistance and offers the ability to set the future direction for existing and future product concepts (with product development managers in insect control) to maximize profit and product longevity. Desired Skill Set MS/PhD in Computer Science/Applied Math/Entomology This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Insect Resistance Modeling Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Spray Automatization Nucleation in Two Phase Holes Project Goal Explore the nucleation of holes that form in liquid sheets that initiate sheet breakup when 2-phase systems are sprayed through agricultural flat fan and air induction nozzles. Background and Motivation This is an extension of prior work that focuses on the mechanistic modeling of the breakup of 2-phase systems when delivered using agricultural nozzles. The development and extension of mathematical models from first principles regarding the breakup of liquid sheets into droplets has been proposed almost 50 years ago but only recently has theoretical work been undertaken to understand this phenomenon. A two-liquid phase (oil-in-water) continuous sheet of fluid emanating from a spray nozzle breaks into drops is first initiated by forming holes within the sheet. Corteva has various models describing hole grow (once formed) and collision to form ligaments, where the ligaments break into droplets via the well-known Rayleigh-Plateau stability theory. What is lacking is a hole perforation model, and thus model refinement is required for the development of a nucleation model that accounts for hole formation within a continuous sheet of fluid that begins this process (which is largely unknown today). Use of analytical/math techniques to quantify experimental observations aligns specifically with the Stats/Math group of Corteva agriscience. Understanding the ultimate drop size distribution that a liquid sheet succumbs to is a capability development for the company and industry in modeling spray atomization. Once nucleation of holes within the sheet is known and adequately addressed, Corteva agriscience has all the fundamental modeling pieces in place for developing the next generation of low drift two-phase formulations to achieve low drift potential [i.e., nucleation of holes (what will be addressed by this proposal), hole growth and collision to form ligaments, followed by ligament breakup into drops]. A strong understanding of fluid dynamics is required. Project Description This is a physical modeling approach via fluid mechanics which begins with an oil-in-water system where the oil phase (e.g., oil drop) is large enough to touch both interfaces of a liquid sheet emanating from a spray nozzle (it is assumed this is what must happen to initiate a hole in a sheet that ultimately leads to sheet rupture and is supported by video observations) which is assumed to lead to a hole forming within a sheet. Holes that form ultimately leads to sheet rupture. The modeling tool(s) to be developed can be a first principles high level programming language (Python) which will be used to supplement the other tools that have been developed once holes are formed. Approaches beginning with hole nucleation, coupled with hole growth, ligament formation, and ligament breakup in drops can coupled to generate a comprehensive tool created from this work. Deliverable(s) and Impact A modeling tool that can be used to recover experimental observations of the breakup of two-phase sheets from agricultural nozzles beginning with nucleation of holes that form in these fluid sheets to ultimately obtain the droplet size we are after (e.g., reduction of drops < 150 um) is the final piece that is lacking for sheet breakup of two-phase systems. This project will focus on the nucleation of holes (where, why and how) Formulation and Application team members will benefit (Brandon Downer, Paul Larsen, Abrin Schmucker) who routinely randomly design and test different formulations and often take the “best” from limited experiments/observations. However, a modeling tool for determining the nucleation of holes within a properly designed 2-phase formulation is sought such that optimal spray characteristics are obtained [e.g., ultimately reduce the driftable fine fraction when applied by standard application methods (e.g., ground and aerial spray) while keeping the larger drops of similar size]. This is a carry-over project from summer 2019 where the student (Jason Turner) took up where the IIT Madras masters student (Nikhil Rajan) left off, refining the python code for hole formation, expansion and breakup into droplets, with refinement on why holes are initiated in two-phase systems (the last unknown in modeling this behavior) Desired Skill Set Fluid mechanics knowledge, Engineering (Mechanical, Civil, Chemical), Theoretical and Applied Mechanics, Applied Math, Applied Physics. This is a carry-over project for Fall 2019 from full time summer work by Jason Turner (MS UIUC Aerospace Engineering) This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Spray Automatization Nucleation in Two Phase Holes Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Ensembles for Demand Sensing Project Goal Algorithms capable of building orthogonal ensembles for predicting units sold by product / district Feature engineering around inventory-based features to detect supply constraints Background and Motivation The business analytics team within DSI is currently leading the Demand Sensing initiative. The DSI team builds the analytical foundation behind this which maintains our customer focus while decreasing inventory write downs. The project will focus on key elements of the initiative around building a repository of models and coming up with different/orthogonal features. As is, this initiative is one of the highest impact/visibility initiatives at Corteva. Project Description Intern will engage in the research and development of algorithms that can forecast unit sales as a function of product attributed and previous sales history. The interns will advance earlier solutions; help build relevant features around inventory and understand the impact of product pricing on demand. The eventual goal is to reduce product write-downs by giving reps and product life cycle managers a data driven tool to better estimate the unit sales by product. Deliverable(s) and Impact What will be delivered at the end of the project: A new set of models/ensembles to predict unit demand Additional features around inventory Understanding the impact of pricing on unit demand Our initiative is cross-enterprise with multiple stake-holders and users including platform leadership. We have an ambitious target of saving the company about 100 MM mainly due to operational effectiveness and product write-downs. If we are able to achieve that, it will lead to significant improvement in ROIC and will have the potential to change the course of our business. Desired Skill Set Experience with Python programming stack is required, as demonstrated by course work, projects (personal or otherwise), previous internships, etc. Some knowledge of econometrics and price/qty relationships preferred but not required This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Ensembles for Demand Sensing Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Predicting protein expression of toxins from different construct designs for insect control Background and Motivation Delivering multiple modes of action (MOA) via molecular stacking necessitates dialing in expression of multiple transgenes between the levels of protein accumulation that are high enough to control insects or provide herbicide tolerance and low enough to avoid negative agronomic effects. Adjusting transgene expression to this “therapeutic window” remains a mostly empirical task that involves iterative testing of various combinations of expression elements (promoters, Expression Modulating Elements, terminators, introns, coding sequences, targeting peptides, etc.). Project Description Datasets: Corn datasets (we currently have Corn 3rd gen that has Cry1B.34, Cry1B.64, and IPDO83Cb) including as many different toxins as possible, with construct design information, gene cassette information and protein expression levels Cytation 5 protoplast data (from Shane Abbit & Sunil Kumar) in which expression modulating elements were used to modify the protein expression levels of toxins. We can leverage this information to adjust protein expression of toxins to a “desired” level. Deliverable(s) and Impact What will be delivered at the end of the project: A model that accurately predicts target protein expression levels given a specific construct design, be it a single, double, or triple gene cassette design What is the impact that the deliverables are going to create: Reduce the number of constructs that need to be tested in stable assays or in the field for Trait Lead Development. This will significantly reduce the costs. Desired Skill Set Experience with machine learning, statistics, deep learning, and databases Ability to write Python/R script Familiarity with different Bioinformatics tools A general understanding of molecular biology and data science concepts This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Predicting protein expression of toxins from different construct designs for insect control Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Interactive Visualization of Synthetic Biology for Insect Control Project Goal This project aims to develop an R Shiny web app for visualizing and interacting with rules/construct designs for different toxins and stack configurations for multi-MOA (mode of action) stack designs. Background and Motivation Delivering multiple modes of action (MOA) via molecular stacking necessitates dialing in expression of multiple transgenes between the levels of protein accumulation that are high enough to control insects or provide herbicide tolerance and low enough to avoid negative agronomic effects. Adjusting transgene expression to this “therapeutic window” remains a mostly empirical task that involves iterative testing of various combinations of expression elements (promoters, Expression Modulating Elements, terminators, introns, coding sequences, targeting peptides, etc.). Project Description Datasets: At least one cohort comprising thousands of insect control stack designs, with gene cassette information and protein expression levels Tools: R environment including R Shiny and ggplot2 or similar for interactive visualization Trains decision trees on-the-fly to generate new design “rules’ given arbitrary therapeutic window constraints Deliverable(s) and Impact What will be delivered at the end of the project: An R Shiny app that: Visualizes toxin expression data for multiple synthetic biology construct designs allows users to include different datasets allows changing the minimum and maximum protein expression levels for different toxins generates and displays different rules What is the impact that the deliverables are going to create: The R Shiny web application will allow enable visual interaction with the data and rules and provide insights and analysis that has not previously been done. Desired Skill Set Experience with machine learning (decision trees and neural networks) and statistics (probability models, standard error, confidence intervals, and linear regression) Ability to write R scripts and develop R Shiny app A general understanding of molecular biology, biotechnology, and synthetic biology This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement .At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Interactive Visualization of Synthetic Biology for Insect Control” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.

Environmental Feature Design for Digital Agriculture Project Goal Improve pedostransfer function models to estimate functional soil attributes at various scales (global, regional, and local) Develop novel features for environmental and ag system modeling, especially features derived from satellite imagery, soils, hydrology and terrain. Build reusable data engineering pipeline components. Background and Motivation High-quality soil and environmental data is critical to enable the expansion of Corteva’s digital service internationally and the global GxExM initiative. As we build out the global soil data system, there are gaps to fill. Although we have good coverage of fundamental soil attributes, the functional soil attributes, e.g., the soil water characteristics, soil profile drainage rate, water table, run-off curve number are generally not available. There is a need to develop reliable and accurate pedotransfer functions to estimate these soil and environmental attributes both at global scale and regional/local scale that fit to specific environment. In addition, there is strong need to continue to improve the accuracy/precision of the existing soil data at scale. Remote sensing science is a viable solution to this problem, which requires engineering new features that help explaining the soil-landscape processes both in space and time. Finally, reusable data pipelines that enable the data science research and predictions will benefit future work. Project Description Research and develop improved and independently validated pedotransfer function models with machine learning to predict functional soil attributes. Drainage class/rate Water table features Curve number Saturated/unsaturated flow characteristics Engineer new features, especially from satellite imagery, that improve capturing the soil-landscape processes both in space and time. Complex soil feature engineering (e.g. NCCPI) Hydrologic soil behavior (e.g. modeled water stress) Write up reports/papers necessary to document the science work including data, methods, results and discussion. Build data pipeline components that funnel data into the machine learning models for model training and prediction. Deliverable(s) and Impact: What will be delivered at the end of the project: Pedostransfer function models to estimate functional soil attributes at various scales (global, regional, and local), e.g., bulk density, field capacity, wilting point. Novel features developed for environmental and ag system modeling, including features derived from satellite imagery. Reports/papers that document methods, validation, Reusable data engineering components that enable automation of discoveries. Who is going to use: Environment and Ag System Modeling group: Xiong Xiong, Naveen Sampath, Dening Ye, Riley McDowell, Brad Malone Remote sensing group: Jeremiah Barrs, Olaniyi Ajadi, Sangzi Liang. What is the impact that the deliverables are going to create: Enable the expansion of digital service in global markets Enable the global GxExM analysis. Improved accuracy of soil/environmental input for crop growth model, I.e., CRONUS Skill set of the desired candidate: Supervised and unsupervised machine learning Geospatial data analysis and remote sensing Python/R, docker Helpful domain background: computer science, soil science, hydrology, environmental science, meteorology, agronomy This role is open to current students of the University of Illinois – Champaign/Urbana campus and is eligible for up to 40 hours of work time per week. Semester: Summer 2020 (May 18 – August 6, 2020) About Corteva Corteva agriscience™ is the only major agriscience company completely dedicated to agriculture. By combining the strengths of DuPont Pioneer, DuPont Crop Protection and Dow AgroSciences, we’ve harnessed agriculture’s brightest minds and expertise gained over two centuries of scientific achievement . At Corteva agriscience™, we are driven by our beliefs and our purpose, which is to enrich the lives of those who produce and those who consume, ensuring progress for generations to come. To apply, email your resume to Lisa Brooks: Lisa.DunganBrooks@corteva.com. Please reference “Environmental Feature Design for Digital Agriculture Internship” in the Subject line and within the email body. Your resume will be forwarded to the proper hiring team for consideration.