I build high-performance systems and the teams that sustain them. After leading cross-functional teams across academic research (HHMI Janelia), startup environments (Elephas Biosciences), and 21 years of military service, I'm transitioning into full-time technical leadership; mentoring the next generation to build production-ready systems with the same rigor and long-term thinking that has kept my open-source tools in use worldwide for 12+ years.
Rare combination of PhD-level technical expertise in GPU-accelerated ML systems and exceptional ability to translate complexity into clarity. I bridge technical teams and product vision, speed development, and build team cohesion through empathetic mentorship.
I focus on sustainable, maintainable solutions over quick wins. My open-source tools have been in continuous development for 12+ years and are still actively used worldwide—because I build systems that serve real workflows, not impressive demos.
I multiply impact through delegation and mentorship. After training personnel in high-stakes military operations, advising 170+ international scientists, and leading diverse software teams, I'm focused on developing technical leaders who combine engineering excellence with outcome-driven discipline.
Keep teams anchored to success criteria at every decision point—build what's needed, not what's possible. Published framework in Journal of Cell Science (2020): "Hypothesis-driven quantitative fluorescence microscopy".
The Challenge: Biological microscopy analysis too slow—days per dataset with traditional CPU methods.
The Solution: Custom CUDA kernels with accessible Python/MATLAB wrappers—leverage GPU acceleration without writing CUDA yourself.
Impact:
- 💯 100x speedups through GPU acceleration
- 🌍 Global adoption by elite research institutions worldwide
- 📚 Published in Bioinformatics (2019)
- ⏱️ 12+ years continuous development (first commit August 2013)
Key Innovation: Accessible interfaces matter as much as raw performance—the best optimization is one people actually use.
Technologies: C++, CUDA, DirectX | Python/MATLAB bindings | CMake, vcpkg
GitHub • MATLAB File Exchange • Publication
The Challenge: 5D microscopy data (x, y, z, channels, time) is incomprehensible when collapsed into 2D projections—researchers draw false conclusions about spatial relationships.
The Solution: GPU-accelerated interactive visualization that exploits human vision strengths (motion detection, depth sensitivity, pattern recognition).
Impact:
- 👥 Used by 170+ scientists at HHMI Janelia and collaborating institutions
- 🔬 Enabled discoveries published in Nature and Nature Communications
- ✅ Prevented algorithmic errors through visual validation with polygon overlay
- 🎬 Real-time rendering with adjustable-speed temporal playback and stereoscopic 3D
Key Innovation: Technology should amplify human capability (visual perception), not replace it.
Technologies: C++ (79%), DirectX, HLSL shaders | MATLAB MEX interface | CMake
The Philosophy: Organize research code when you have 10 functions, not 1,000—modularity compounds, each reusable function becomes foundation for the next.
The Solution: Structured collection of microscopy analysis utilities treating research code like production software from day one.
Impact:
- 🏗️ Foundation for Hydra Image Processor and Direct 5D Viewer
- 🌍 Used globally at research institutions and collaborating laboratories
- 📅 11+ years continuous development (started 2013)
- 📖 "Show your warts" — Public development showing evolution, mistakes, and learning moments
Key Innovation: Unified reader/writer system for all microscope formats—write analysis code once, apply to any microscope data.
Technologies: MATLAB packages | JSON metadata | Multi-format support (TIFF, OME-TIFF, CZI, ND2)
The Challenge: Finding "great photos" in massive personal collections—combining technical metrics (sharpness, exposure) with subjective aesthetic judgment.
The Solution: Computer vision + YOLO object detection + hierarchical classification to surface the best images from 200,000+ photo libraries.
Impact:
- 🎯 10x efficiency gain — Review 2,000 candidates instead of 20,000 images
- 🗄️ Scale-tested on 700K+ file collections
- 🌲 Flexible taxonomies — Biological classification (World Flora Online), geographic hierarchies, custom categories
- 🔒 Long-term reliability — Checksummed tracking, never moves/deletes originals
Key Innovation: ML should multiply human capability, not replace it—surface the best candidates, let humans make final decisions.
Technologies: Python, SQLAlchemy, MariaDB | OpenCV, YOLO, scikit-learn | Jupyter Lab
All major projects publicly available and actively maintained:
- 12+ years maintaining Hydra Image Processor (since 2013)
- Open development showing evolution, mistakes, and learning moments
- Accessible interfaces considered alongside raw performance
- Global adoption by research institutions and collaborative teams worldwide
- Reproducible science through documented pipelines and version control
I'm actively seeking technical leadership roles where I can shape engineering culture, mentor teams, and build production ML systems that matter.
Open to:
- Staff/Principal Engineer → Engineering Manager transitions
- Technical Lead roles with mentorship responsibilities
- Director of Engineering positions at mission-driven organizations
- CTO/VP Engineering roles at startup organizations
Ideal environments:
- Mission-driven organizations where technical excellence serves meaningful purpose
- Teams that value sustainable engineering practices and long-term thinking
- Cross-functional collaboration between technical and business stakeholders
- Mentorship and leadership development culture
📧 Email: info@ericwait.com 💼 LinkedIn: linkedin.com/in/ericwaitinfo 🌐 Website: ericwait.com 📄 Resume: ericwait.com/pdfs/Eric_Wait.pdf
"The camera is an instrument that teaches people how to see without a camera." — Dorothea Lange
My professional photography training taught me pattern recognition that informs data visualization, compositional discipline that ensures reproducible systems, and a fundamental truth: the best solution often requires subtracting, not adding—whether in photographs, ML architectures, or team processes.