Education of students, post-docs, and faculty on programming and Best Practices in Software Development is a large part of MolSSI's mission. Our education program consists of our cohorts of Software Fellows, online training materials, and multiple workshops online and at various locations each year around the US.
The MolSSI continues to fund prestigious software fellowships that recognize advanced graduate students and postdocs pursuing the development of software infrastructure, middleware, and frameworks that will benefit the broader field of computational molecular sciences, including biomolecular and macromolecular simulation, quantum chemistry, and materials science.
The journey into molecular science for Shuhang Li, a current PhD candidate at Emory University, began in high school where he was inspired by two extraordinary teachers: Mrs. Geng, a Chemistry teacher, and Mr. Du, a Physics teacher. Mrs. Geng had a unique talent for transforming abstract chemical concepts into relatable stories that made molecular interactions feel tangible. At the same time, Mr. Du’s lessons in physics revealed the fundamental forces that govern matter at every scale, from quantum interactions to celestial mechanics. Their passion and guidance made him appreciate the potential of natural science as a pursuit, and this curiosity led Shuhang to pursue deeper study at the College of Chemistry at Nankai University.
Bridging Mathematics, Physics, and Computing
During his tenure as a MolSSI Software Fellow, Shuhang has had the privilege of being mentored by Dr. Jonathan Moussa, a MolSSI Software Scientist with expertise spanning mathematics, physics, and computer science. His enthusiasm for problem-solving was both inspiring and transformative. One pivotal moment in research came when Shuhang was struggling with optimizing software performance due to computational bottlenecks. Dr. Moussa introduced the concept of runtime code generation—an elegant solution where code is dynamically generated during execution rather than explicitly written and stored. Additionally, he suggested an approach to optimize tensor operations by strategically generating intermediate tensors, significantly enhancing computational efficiency. These insights revolutionized the approach to software development, demonstrating the power of mathematical intuition in crafting high-performance computational tools.
Current Research
As a MolSSI Software Fellow, Shuhang focuses on designing robust and efficient multireference electronic structure algorithms to simulate electronic excitation, ionization, and attachment processes in open-shell systems. These species, such as radicals and biradicals, play a crucial role in photochemical reactions and electron transfer, making their electronic structures essential for understanding chemical mechanisms.
A key aspect of this work involves developing NiuPy (GitHub link), a software that integrates wicked (GitHub link), an automated algebraic derivation tool, with a runtime code generator that streamlines tensor contractions. By automating labor-intensive steps, this framework eliminates manual implementation bottlenecks, enabling rapid prototyping for researchers working on multireference equation-of-motion (EOM) theories and quantum computing approaches.
As a Ph.D. student in the Evangelista Lab at Emory, Shuhang is also developing low-scaling multireference electronic structure algorithms to overcome computational bottlenecks when modeling large systems or active spaces. These efforts aim to extend the applicability of existing multireference renormalization group methods, making them viable for previously intractable problems.
The Future: Open Science and Academic Pursuits
Through the MolSSI Fellowship, Shuhang has gained transformative skills that bridge computational theory with real-world software development. Beyond coding, this experience has underscored the power of open-source software in accelerating discoveries across physics, chemistry, and materials science. The long-term goal is to remain in academia, contributing to the development of powerful computational tools that bridge disciplines and empower researchers worldwide.
Beyond Research: Cooking, Hiking, and a Dream of Tibet
While science is a driving force, Shuhang also enjoys exploring the hiking trails of North Georgia (and beyond!) whenever time permits. Cooking, particularly Cantonese cuisine, has also become a favorite pastime pursuit.
A significant personal milestone for Shuhang is becoming the first Ph.D. in his family. Though the journey has been rigorous, it has also been deeply rewarding. Looking ahead, one personal dream remains—to travel to Tibet, a place of breathtaking landscapes, rich culture, and deep history. For now, however, the focus remains on pushing the boundaries of computational science. But beyond the equations, algorithms, and simulations, Shuhang Li remains committed to a broader vision—one where science is open, collaborative, and deeply connected to the world.
You can find more about him at LinkedIn | Github | X (former Twitter) @Li_Shuhang
Levi Petix’s path to molecular science was not initially planned. When he began graduate school, he was open to exploring various research avenues within chemical engineering. However, during the first semester, faculty members presented their research to recruit new students, and it was during this time that Levi discovered his passion for molecular science. Prof. Michael Howard’s presentation stood out, showcasing the importance of computational research in soft materials beyond just scientific curiosity. Levi was intrigued by how molecular science enables the rapid identification and screening of solutions to complex research challenges while minimizing time and costs associated with extensive experiments. Despite his deep dive into this field, he maintains an engineering mindset, always focused on practicality and application. He finds immense satisfaction in leveraging fundamental physics and computation to tackle difficult yet meaningful problems.
Current Research
Levi’s current research as a MolSSI Software Fellow focuses on multiscale inverse design, with an increasing emphasis on methods development. He’s particularly interested in leveraging computational techniques to bridge the gap between simulation and experiment. As he notes, material discovery has traditionally relied on experimental forward design, a process that involves selecting chemicals, conducting experiments, and characterizing the resulting structures. If the desired material properties are not achieved, researchers adjust experimental conditions and repeat the process—a costly, time-consuming, and labor-intensive cycle. However, advances in computational techniques are now enabling a more efficient approach: multiscale inverse design.
As a Ph.D. candidate at Auburn University, Levi is at the forefront of this field, focusing on relative entropy minimization as a key technique for inverse design. Unlike black-box machine learning or iterative Boltzmann inversion, this approach allows for the direct incorporation of physically meaningful constraints, making experimental validation more feasible. Originally pioneered by Prof. Scott Shell (UCSB) and later by Prof. Tom Truskett (UT Austin), relative entropy minimization has been highly effective in designing coarse-grained models. However, past implementations primarily focused on unconstrained pair potentials, limiting direct experimental applicability.
Through his MolSSI fellowship, Levi is expanding relentless, a computational framework designed to model and optimize materials. This expansion enables the application of relative entropy minimization to materials with bonds, such as polymers and metal-organic frameworks, making it a powerful tool for linking simulations with experimental synthesis. The overarching goal is to leverage high-performance computing (HPC) to rapidly screen material designs, eventually providing experimentalists with precise recommendations for synthesis.
Advancements in Methodology and Software
Levi has integrated surrogate modeling into the relative entropy framework, significantly accelerating the optimization of interaction parameters. In collaboration with Prof. Chris Kieslich (Georgia Tech), this approach has reduced the sequential nature of parameter optimization, enabling parallel training of surrogate models and shortening the time to solution. Their work was recently published in the Journal of Chemical Theory and Computation (doi.org/10.1021/acs.jctc.3c00651).
In addition to surrogate modeling, Levi collaborates with Prof. Jeetain Mittal’s group (Texas A&M) on physics-informed inverse design, incorporating physical constraints directly into parameter optimization in relentless. This approach enhances experimental relevance and improves simulation efficiency.
Additionally, a collaboration with Prof. Tom Truskett’s group (UT Austin) focuses on applying relentless to coarse-grain polymer systems, dramatically accelerating calculations.
On the software development side, Levi is the lead developer of relentless and a co-developer of lammpsio, a package that simplifies LAMMPS data file creation and conversion to HOOMD-blue’s GSD format. Contributions have also been made to AZplugins, a widely used HOOMD-blue plugin maintained by Prof. Michael Howard and Prof. Antonia Statt (UIUC), and Jupyter notebook tutorials for HOOMD-blue’s Multiparticle Collision Dynamics (MPCD) module, in collaboration with Prof. Jeremy Palmer’s (University of Houston) group.
Towards an Integrated Multiscale Workflow
Beyond mesoscale modeling, Levi conducts atomistic simulations to investigate solvent-mediated interactions. With the ongoing expansion of relentless, the framework will soon enable coarse-graining of polymers from atomistic simulations, providing a direct connection between molecular-level interactions and mesoscale material properties. His long-term vision is to integrate relentless into a broader computational pipeline, where atomistic simulations inform mesoscale modeling, ultimately leading to materials that can be experimentally synthesized. By bridging computational predictions with laboratory realization, Levi aims to revolutionize materials design, making it more efficient, predictive, and experimentally relevant.
The Influence of his Software Sciences Mentor
Levi credits much of his development to his MolSSI software sciences mentor, Dr. Jing Chen, whose expertise in both polymer science and computer science has been invaluable. Her insights have played a crucial role in shaping relentless, a computational tool designed for efficient and transparent soft material design. With Jing’s guidance, Levi has been able to refine the software’s design and layout, ensuring its effectiveness and usability. As anyone in software development knows, code that runs smoothly on one computer often encounters issues on another, and Jing’s willingness to test and troubleshoot has been instrumental in overcoming these challenges.
The Impact of the MolSSI Software Fellowship
Being awarded the MolSSI Software Fellowship has been a transformative experience for Levi. The fellowship has provided him with the opportunity to expand relentless in ways that will greatly benefit the computational soft material research community. Previously, relentless was limited to designing materials that interacted through pairwise potentials without bonds. However, with the support of MolSSI, Levi has been able to incorporate molecular materials with bonds, opening up new research possibilities for addressing long-standing questions in his research group and the broader scientific community.
Beyond the technical advancements, the fellowship has been a period of tremendous professional growth for Levi. His work with relentless has transitioned from minor bug fixes and small feature additions to a large-scale expansion, requiring careful planning, a well-structured development roadmap, and a strong focus on user interface, documentation, and testing. These experiences have enhanced his skills as a software developer and will be invaluable in his future.
Future Career Aspirations
Looking ahead, Levi envisions a career that bridges industry and academia. His immediate goal after graduate school is to work as a computational scientist in the industry, where he hopes to apply his
expertise to real-world challenges. However, his long-term ambition is to return to academia as a professor of practice. Inspired by Professors of Practice David Carroll and Mike Gill at the University of Mississippi who shaped his development as an engineer and leader, Levi aspires to mentor and educate future generations of engineering students by combining his industrial experience with the teaching skills honed during graduate school.
Unique Talents and Passions
Outside of his academic and professional pursuits, Levi has an uncanny knack for securing incredible deals on event tickets. Whether it’s concerts or sporting events, he has managed to see artists like Noah Kahan, Lucy Dacus, Taylor Swift, and Fleetwood Mac up close for minimal cost. His most remarkable ticketing feat occurred during the AIChE 2022 conference in Phoenix, where he found center-stage tickets for an Elton John concert just two hours before the show for only thirty dollars each, allowing him and his lab mate to witness one of Elton’s final performances.
Proud Accomplishments
Among his many achievements, Levi considers receiving the MolSSI Software Fellowship his most significant academic accomplishment. The fellowship has provided him with the opportunity to collaborate with an exceptional group of researchers and developers, enhancing his skills and deepening his impact on computational molecular science. It has been an invaluable experience that continues to shape his career.
Life Beyond Research
When he’s not working, Levi is an avid sports fan. He dedicates much of his free time to following and attending games, particularly those of his alma mater, Ole Miss, as well as Auburn football and baseball. He also dreams of traveling to London to watch his favorite Premier League team, Tottenham Hotspur, play after completing his Ph.D. His love for sports and travel has led him to explore various stadiums, including Virginia Tech’s Lane Stadium and English Field during his visit to MolSSI, where he even had the unique opportunity to step inside Lane Stadium for the attached photo!
Immediate Goals
The upcoming years promise to be busy and exciting for Levi. He is preparing for a major personal milestone—his wedding in March—and is also focused on completing his Ph.D. by the spring of 2026. His immediate goals include publishing research papers using the enhancements made to relentless with MolSSI’s support, wrapping up other ongoing projects, and securing a job in industry. Balancing these ambitious endeavors while maintaining a healthy work-life balance is a priority, and Levi is determined to make the most of this critical phase of his journey.
Find more about him on GitHub, LinkedIn and X (former Twitter @LeviPetix)
The Molecular Sciences Software Institute (MolSSI) and Intel invite you to attend our hands-on Basics ofAccelerated Computing with Intel OpenMP GPU Offload virtual workshop on Mar. 7, 2025 from 13:00 – 17:00 ET.
The main focus of the course will be on the following topics:
A brief overview of OpenMP parallelization on CPUs
Offloading an OpenMP CPU-parallel code to GPUs
Managing data transfer between the host and the device within OpenMP framework
OpenMP GPU offload best practices
Overview of a real-life example in which Intel OpenMP helped community a code such as NWChem to perform compute- and data-intensive tasks more efficiently
The following prerequisites are recommended but not mandatory for attending the workshop:
Experience with OpenMP parallelization on CPUs
Basic familiarity with Bash as well as C, C++ or Fortran programming languages
Familiarity with Jupyter Lab environment
We have arranged a brief preparation session before the beginning of the workshop. The details of the preparation session can be found here.
You can stay up-to-date about the upcoming workshops by subscribing to our newsletter here or following the updates posted on our Industrial Training Program webpage.
The Molecular Sciences Software Institute (MolSSI) and Intel invite you to attend our hands-on Introduction to Intel® Tiber™ AI Cloud and oneAPI virtual workshop on Mar. 7, 2025 from 12:00 – 13:00 ET.
The main focus of the course will be on the following topics:
Intel Certified Instructor program
Introduction to Intel oneAPI ecosystem
Intel Tiber AI Cloud, registration and basic usage
Software, hardware and educational resources available on Tiber AI Cloud
Hands-on example: Running a conversational app in Jupyter Lab based on DeepSeek-R1 large language model
The following prerequisites are recommended but not mandatory for attending the workshop:
Experience with cloud environments (Oracle OCI, AWS, Microsoft Azure, Intel Tiber AI Cloud etc.)
Basic familiarity with Bash as well as Python programming language
You can stay up-to-date about the upcoming workshops by subscribing to our newsletter here or following the updates posted on our Industrial Training Program webpage.
Caitlin Whitter didn’t take the traditional path into molecular science. As a Ph.D. student in computer science at Purdue University, she always knew she wanted to work on interdisciplinary research. Rather than sticking purely to computing, she was drawn to solving complex problems in the natural sciences. That curiosity led her to computational molecular science, where she now applies machine learning to better understand molecular and atomic properties.
Diving Into Computational Science Whitter’s research focuses on building interpretable machine learning models to predict molecular and atomic properties accurately and efficiently using quantum-mechanical datasets. These predictions have significant implications for fields like drug discovery, materials science, and chemistry, where understanding molecular interactions is crucial. By developing interpretable machine learning pipelines, she aims to make these computational tools more effective and accessible to scientists.
Identifying molecular compounds with novel properties is a vital step in pharmaceutical drug discovery and materials design. However, the chemical space is vast, with approximately 10^60 possible molecules, creating a strong need for accurate and efficient predictive models. Traditional quantum chemical methods, such as Density Functional Theory (DFT), can take days per molecule to achieve high accuracy. Machine learning models, on the other hand, offer a way to dramatically accelerate this process, predicting properties in mere seconds. Yet, the lack of transparency in typical black-box models can hinder scientific trust and usability.
To address this, Whitter designs machine learning pipelines with interpretability in mind, ensuring clearer insights into the underlying decision-making process. During her Ph.D., she has worked on multiple projects at the intersection of machine learning and computational chemistry. These include:
Physics-informed graph convolutional networks for fast and accurate prediction of atomic multipole moments.
Subset selection algorithms for choosing representative molecular science datasets to improve neural network performance and gain insights into dataset distributions—a focus of her MolSSI software project.
The MolSSI Fellowship Experience Becoming a MolSSI Software Fellow has been a transformative experience for Whitter. The fellowship provided funding that allowed her to fully dedicate herself to her MolSSI project for a year, diving deeper into her research without financial concerns. One of the highlights has been working closely with her MolSSI Software Scientist mentor, Dr. Benjamin Pritchard. Their bi-weekly meetings have provided invaluable insights, particularly in working with molecular science datasets. The MolSSI summer bootcamp was another major advantage—not only was it a great learning opportunity, but it also allowed her to connect with other Fellows and explore research from different perspectives.
Developing Machine Learning for Molecular Science In her MolSSI software project, Whitter focuses on developing subset selection algorithms to improve the training efficiency of neural networks while maintaining low error. These algorithms also provide deeper insights into the distribution of molecular datasets. One of the benchmark datasets she works with is QM9, a quantum-mechanical dataset containing 134,000 small molecules with various energetic and electronic properties. This dataset is widely used in computational chemistry and is available through multiple repositories, including MolSSI’s QCArchive.
To analyze these molecular datasets, she trains graph neural networks (GNNs), a type of neural network specifically designed for graph-structured data. Because molecules are often represented in tabular form, she first converts them into graph representations, where atoms are nodes and bonds are edges. Additional molecular features, such as atomic number and bond type, further refine these graph structures.
For implementation, Whitter primarily uses Python and leverages software packages like PyTorch and Deep Graph Library (DGL) for model architecture and training. Additionally, RDKit provides access to molecular features beyond what is included in standard molecular science datasets.
Looking Ahead As she works toward completing her Ph.D., Whitter remains excited about the future of machine learning in computational science. Whether in academia, industry, or national labs, she hopes to contribute to cutting-edge advancements in the field. Beyond research, she enjoys singing in a musical ensemble at Purdue, reading, exploring new places, and spending time with friends and family. Looking back, she is proud of the collaborations she has been part of and the opportunities she has had to share her research with a broader audience. Being selected as a MolSSI Software Fellow has been one of the most rewarding experiences of her Ph.D. journey, and she highly encourages others interested in computational molecular science to apply.
To connect with Caitlin Whitter or learn more about her research, visit her LinkedIn profile
Dates: July 9-10, 2025
Location: Cal Poly San Luis Obispo, San Luis Obispo, CA
This workshop is designed for university faculty in the molecular sciences (chemistry, physics, molecular biology, materials science, and related disciplines) who want to learn Python programming and how to incorporate it into their class. This workshop targets faculty who have no or very limited python programming experience and want to upskill and then incorporate cyberinfrastructure skills, like programming, data analysis, data visualization into their courses. The workshop is open to faculty at all academic ranks, including tenure-stream and instructional faculty.
Workshop Details
The workshop will be held in-person on the campus of California Polytechnic State University in San Luis Obispo, California all-day on Wednesday, July 9 and Thursday, July 10, 2025. We recommend that all participants travel to San Luis Obispo on Tuesday, July 8 to be prepared to participate in the workshop starting at 9:00 am on Wednesday. The workshop will conclude at 5:00 pm on Thursday, July 10. For participants who are traveling from out-of-state, we recommend arranging your flight home on Friday, July 11.
Workshop Topics
Wednesday, July 9: Introduction to data analysis and visualization. Topics covered include:
Basic Python syntax
Reading and writing data from files
Analyzing data with the Python libraries Numpy and pandas
Plotting data with matplotlib
Thursday, July 10:
Faculty will be able to choose from several specific curriculum workshops where MolSSI ACT-CMS Faculty Fellows present curricular activities that use Python in molecular science courses. There will also be sessions about computational platforms that can be used for class actitivites and best practices for teaching programming.
Travel Awards
We’re committed to making this workshop accessible! Apply for one of two travel awards (Participants should indicate what type of travel award they wish to apply for on their application. Self-funded applicants are also welcome):
Full travel award: This award includes support for airfare (up to $600) and three nights hotel accommodation (Tuesday, Wednesday, and Thursday) in San Luis Obispo. Breakfast is available at the hotel each day and lunch is provided at the workshop. Dinner is not provided. These travel awards are targeted for out-of-state participants. 10 awards available.
Mini travel award: This award includes support for two nights hotel accommodations (Tuesday and Wednesday) in San Luis Obispo. Breakfast is available at the hotel each day and lunch will be provided at the workshop. This award does not include reimbursement for mileage or other travel. These awards are targeted for in-state participants. 15 awards available.
To apply for the workshop, please fill out the online application. The application asks for contact information, information about your institution, and a short statement of interest explaining why you want to attend the workshop. If you have questions, you can email us at act-cms@molssi.org.
MolSSI Workshops
The MolSSI’s Software Workshop program is a community-driven effort in which researchers from academia, industry, and national labs propose timely and important topics focused on the software needs of the molecular sciences, and the MolSSI organizes or facilitates the event.
