Software
Research Software
ssbc
PAC-Conformal Prediction with Small Sample Guarantees
Python package implementing Small Sample Beta Correction for conformal prediction. Provides tight PAC coverage guarantees even in small-sample regimes, enabling accountable automation for scientific applications.
Features:
- Distribution-free, finite-sample guarantees
- Operational rate estimation
- Applications in toxicity screening, molecular property prediction, structural biology
- Integration with scikit-learn pipelines
dlsia
Deep Learning for Scientific Image Analysis
Comprehensive ML toolkit for scientific imaging applications including segmentation, anomaly detection, and tensor processing across multiple modalities.
Applications:
- Tomography reconstruction and analysis
- Lattice light sheet microscopy
- Electron microscopy
- X-ray scattering data analysis
Key capabilities:
- CNN-based segmentation
- Anomaly detection algorithms
- Multi-modal image processing
- Integration with experimental workflows
qlty
Out-of-Core Tensor Management
Framework for efficient ML training and inference on memory-constrained hardware. Enables processing of large-scale imaging datasets that exceed available RAM.
Features:
- Intelligent chunking and caching strategies
- PyTorch integration
- Memory-efficient data loading
- Support for distributed computing
Major Infrastructure Contributions
PHENIX
Macromolecular Structure Determination Suite
Core developer and contributor to PHENIX, a comprehensive system for automated determination of macromolecular structures using X-ray crystallography and other methods.
My contributions:
- Xtriage: Quality control and pathology detection
- Twinning detection algorithms
- Integration of computational crystallography methods
- Automation workflows
Impact: Used by thousands of structural biologists worldwide; cited >10,000 times
PHENIX Website
Adams et al. (2010) Acta Cryst. D
CCTBX
Computational Crystallography Toolbox
Foundational crystallographic infrastructure providing core algorithms and data structures for crystallographic computing.
Contributions:
- Data pathology detection algorithms
- Statistical analysis tools
- Integration with experimental beamline systems
- Python bindings and user interfaces
Impact: Forms the computational backbone for PHENIX and numerous other crystallographic software packages
Xtriage
Crystallographic Data Quality Control
Automated quality control system for crystallographic data, detecting common pathologies and data collection problems.
Features:
- Twinning detection
- Ice ring identification
- Wilson plot analysis
- Anomalous signal detection
- Automated reporting
Adoption: Integrated into the Protein Data Bank validation pipeline, checking all deposited structures worldwide
Domain-Specific Tools
SAXS Analysis Pipeline
Small-Angle X-ray Scattering
Developed automated analysis pipeline achieving 10× speedup for SAXS data processing at the Advanced Light Source.
Capabilities:
- Automated data reduction
- Buffer subtraction
- Guinier analysis
- Distance distribution functions
- Integration with experimental control systems
FEL Workflows
Free-Electron Laser Data Processing
Exascale computational workflows for serial femtosecond crystallography and other FEL techniques.
Features:
- Real-time data processing
- Hit finding and indexing
- Structure factor extraction
- Scalable to exascale systems
- Integration with NERSC infrastructure
Development Practices
All research software follows:
- Version control with Git
- Continuous integration testing
- Comprehensive documentation
- Open-source licensing (typically BSD-3-Clause or MIT)
- Active community engagement
Collaborative Development
I actively collaborate on software development with:
- DOE facility computational staff
- Academic research groups
- Open-source community contributors
- Industry partners
Technical Stack
Languages: Python, C++
ML Frameworks: PyTorch, NumPy, SciPy, scikit-learn, scikit-image
Computing: HPC/Exascale systems, CUDA, distributed computing
Tools: Git, GitHub, CI/CD, Docker, Jupyter
Domains: Scientific imaging, structural biology, FEL science, spectroscopy
For publications describing these tools and methods, see the Publications page.