Denny Schaedig — home

Experience

Machine learning engineering across production ML systems, generative models, and research infrastructure.

Professional experience

Laboratory for Child Brain Development — Washington University in St. Louis

St. Louis, MO

Machine Learning / AI Engineer

Dec 2022 – Present

  • Architected a HIPAA-compliant AWS HPC environment (EC2, EFS, S3, Glacier) delivered ~32% under projected cost, and secured an additional 24% university-wide cloud discount
  • Designed and shipped HRfunc, an open-source, pip-installable Python package for signal deconvolution; published in Neurophotonics (2025), presented in an invited NIRx vendor webinar, and maintained across 22 public releases (~14K PyPI downloads)
  • Trained NIRSgen, a diffusion model trained from scratch on biomedical time-series data for data augmentation; validated via transfer learning, raising downstream classification accuracy to 92% (vs. 76% with random reinitialization); currently being prepared for publication, with the model to be released afterward
  • Built BOLDnet, a TensorFlow framework with custom spatial-attention layers and 4D volumetric data loaders for rapid model prototyping on standardized datasets
  • Developing synchronAI, a multimodal deep-learning model that fuses video, audio, and time-series signals for automated inference (0.72 AUC)

Rocky Mountain Digerati, LLC

Summit County, CO

Founder

Nov 2025 – Present

  • Built a healthcare-finance analyst agent with a novel hallucination detection and removal layer, running on open-weight models as small as 16B parameters
  • Operate a heterogeneous multi-node inference cluster (Ray, mixed CUDA and Apple Silicon backends, Tailscale mesh) for local training, evaluation, and model serving
  • Curated and published a 4,040-image computer-vision dataset on Hugging Face (DOI 10.57967/hf/6143; peak of ~2,600 monthly downloads) for avalanche-risk modeling
  • Generatively pre-trained snowGAN (PyTorch / TensorFlow) on high-resolution snowpack imagery, released on Hugging Face (RMDig/snowGAN-core and RMDig/snowGAN-magnified-profile) and transferred into a downstream multimodal avalanche-risk classifier (AvAI)
  • Built the AvAI cross-platform mobile app to serve real-time model inference and collect labeled field data
  • Developed AvServ, a federated server that orchestrates the check-out safety program, connecting AvAI users directly to search and rescue teams when they fail to check in at the expected time

Čeko Lab (The Pain Lab) — University of Colorado Boulder

Boulder, CO

Software / ML Engineer

Jul 2020 – Feb 2023

  • Trained an RNN to predict pain levels from wearable-sensor and questionnaire data (79% accuracy vs. a 63% majority-class baseline)
  • Built ETL pipelines that moved and preprocessed data from remote servers and devices into a relational SQL database, time-aligning multiple physiological streams for ML training
  • Automated the study enrollment process to streamline onboarding of new participants

Intermountain Neuroimaging Consortium — University of Colorado Boulder

Boulder, CO

Software Engineer

Jul 2017 – Feb 2022

  • Built ETL pipelines and automation for MRI scheduling, study enrollment, and training compliance, improving recruitment throughput and data availability
  • Created an automated system to track MRI users' training compliance and a scheduling assistant for finding available scanner time

D'Mello Lab — University of Colorado Boulder

Boulder, CO

Professional Research Assistant

Jan 2020 – Jan 2021

  • Collected, preprocessed, and analyzed neuroimaging data recorded during a reading task to study how readers' attention to text changes across various pathologies
  • Explored algorithms to correct eye-tracking drift artifacts

Banich Lab — University of Colorado Boulder

Boulder, CO

Research Assistant

Apr 2017 – Dec 2019

  • Ran fMRI studies and collected, preprocessed, and analyzed fMRI data using machine learning and standard statistical models
  • First developed BOLDnet (originally named KleinNet), a deep-learning framework for neuroimaging data
  • Developed the software for a working-memory experiment examining how faces and locations transition into short-term memory

Additional experience: Ski instructor at Eldora and Loveland ski resorts, teaching skiers of all ages with a focus on children ages 3–14; first responder and medical officer with the Aurora Medical Team, responding to medical emergencies and teaching CPR and first aid.

Skills

Languages
Python, Go, TypeScript / JavaScript, SQL, Dart, C++, Bash
Machine Learning & Deep Learning
Generative AI (GANs, diffusion models), Transformers / attention, CNNs, RNNs, computer vision (object detection), speech & audio (Whisper, TTS), multimodal fusion, reinforcement learning, transfer learning
LLMs & Agents
Agentic AI, RAG (pgvector, sentence-transformers), hallucination detection & LLM evaluation, local inference (llama.cpp, CTranslate2), Anthropic & OpenAI APIs, LangChain
Frameworks & Libraries
PyTorch, TensorFlow / Keras, Hugging Face Transformers, Ultralytics YOLO, MMDetection, OpenCV, scikit-learn, XGBoost, SciPy, statsmodels, pandas, spaCy
Model Distribution & Serving
Hugging Face Hub (models & datasets), TensorFlow Lite (on-device), ONNX Runtime, vLLM, Ollama, FastAPI inference APIs
Code Distribution & CI/CD
PyPI packaging (hrfunc, pymaxim), GitHub Actions, pytest, ruff, mypy, uv
Infrastructure & MLOps
AWS (EC2, Lambda, S3, EFS, Glacier), Docker / Compose, Kubernetes, Ray, CUDA, Weights & Biases, PostgreSQL, Redis, Prometheus, OpenTelemetry, Cloudflare, Tailscale, Vercel
Web & Apps
FastAPI, Flask, React, Next.js, Astro, Flutter
Neuroimaging & Signals
MNE / MNE-NIRS, Nilearn, NiBabel, NeuroKit2, fMRI, fNIRS, physiological signals

Education

Denny Schaedig at his 2025 M.S. graduation

M.S., Artificial Intelligence & Machine Learning — Colorado State University Global

2025

  • Capstone: applied generative AI to avalanche-risk prediction; snowGAN, generatively pre-trained on snowpack imagery, became the foundation for AvAI

B.A., Neuroscience — University of Colorado Boulder

2019

  • Focus areas: memory and learning, decision making, computer science, and machine learning
  • President of the Neuroscience Club for two years, organizing events and trips

Certifications & Awards

  • Stanford Online — Algorithms: Divide and Conquer, Sorting and Searching, and Randomized Algorithms; Graph Search, Shortest Paths, and Data Structures
  • President's Volunteer Service Award — Corporation for National and Community Service; awarded for 500+ hours of volunteer service, including with the Aurora Medical Team and an Alzheimer's care home
  • Ski instruction: Alpine Level 1 and Children's Specialist 1 certifications, Professional Ski Instructors of America (PSIA); Child Ski Instructor of the Year and Ski Instructor of the Month, Eldora Ski Resort