MONICA SPISAR, PhD machine learning & software engineer | physicist
[email protected] monicaspisar.com . github.com/msyvr . linkedin.com/in/monicaspisar

CORE SOFTWARE SKILLS

Languages: Python, Go, Rust. Data: SQL, PostgreSQL, BigQuery, Rockset. Open Telemetry. Superset, Honeycomb, Datadog, Grafana, Prometheus. DevOps/Infrastructure: Docker, Terraform. Networking: Tailscale, Wireguard. Cloud: GCP, AWS. Basics: Git. Bash. Javascript, HTML, CSS.

Machine learning & data science: PyTorch (+Vision/Audio), NumPy, SciPy, MATLAB + signal/image processing & FEA toolboxes. Graduate level coursework in probability and statistics and in signal and image processing.

TECHNICAL LEAD, INDIVIDUAL CONTRIBUTOR

Software engineer: upskilling & craftsmanship @ Recurse Center 2021 & 2024
Dramatically improved my software engineering skills in both batches at `the writing residency for programmers`.
2024: Machine learning (deep learning models), AI safety (mechanistic interpretability), memory-safe languages (Rust).
Designing neural networks: zero to micrograd, Rust: Notes on memory management. GitHub: micrograd-python, ruray.
2021: Computer science fundamentals. Backend engineering focus, mainly in Python, Go. Code craftsmanship.
Naive ray tracer implementation in Python. GitHub: raytracer (Python).

Software engineer: censorship circumvention systems @ Lantern 2022 - 2024
Member of a technical team managing distributed cloud infrastructure serving millions of users. Designed and rebuilt a streaming data pipeline, migrated a data warehouse: reduced data storage and processing costs by 50%. Helped reduce service outage durations up to 10x via data monitoring. Designed custom metrics to support dev, biz, and client services teams. Go, Python, Rust, GCP, Docker, Terraform, Open Telemetry, Superset, Big Query, Honeycomb, Datadog, Tailscale.

Research engineer: medical devices @ Kardium (Employee #16) 2008 - 2011
Led deployment imaging for a class III medical device for transcatheter mitral valve repair. Led device performance characterization (computer simulations, lab), preclinical trial design, initial clinical evaluation for a class II device for sternal closure. Patents: 8888791, 9700363: Surgical instrument and method for tensioning and securing a flexible suture

Research scientist: biomedical imaging lab @ Sorbonne University (Postdoc) 2003 - 2004
Computer simulations of fluid and small particle flow in human microvasculature-like structures. Designed, built a microfluidics vascular flow prototype for high resolution ultrasound imaging with contrast agents.

Research assistant: biomedical ultrasound lab @ University of Michigan (PhD) 1998 - 2003
Designed/built a laser-based ultrasound imaging system. Novel detection technology. Met clinical specs, 10x sensitivity increase. Signal capture, image processing, image reconstruction, systems analysis. Pre-thesis research: Built a small scintillation (gamma) camera for breast imaging. Monte Carlo simulations, statistical image reconstruction methods.

PERSONAL SOFTWARE PROJECTS, samples from github.com/msyvr

Machine learning & agents: micrograd implementation, mechanistic interpretability, agentrix, model evaluation
Ray & path tracing engines: Rust, Python

PERSONAL BLOG, samples from monicaspisar.com

Designing neural networks: zero to micrograd, Rust: Notes on memory management, RAG + fine tuning, Herding processes: async + parallel, Evaluating machine learning models, Bayesian persuasion perspective on AI safety

LEADERSHIP, MANAGEMENT, ENTREPRENEURSHIP - translational skills essential for complex technical projects

Portfolio Manager / Scientific Liaison @ University of Oxford (2019 - 2021), COO @ Mineral Deposit Research Unit (2013 - 2015), Program Manager & Industry Grants Officer @ University of British Columbia (2011 - 2013), Founder/CEO @ Little Stars (2010 - 2014), CEO @ Panne Rizo (2004 - 2011), Startup team member @ Xoran Technologies (2000 - 2001)

EDUCATION + RESEARCH TRAINING

PhD, Biomedical Engineering: Medical Imaging, University of Michigan
Thesis: Optoacoustic detector arrays for medical imaging applications. Publications: Google Scholar, ResearchGate

BSc, Physics, University of Toronto