Matt M.
ML Audits, Prediction Improvement & Forecasting | Kaggle Grandmaster
Is your model underperforming, your validation unreliable, or your forecasting pipeline failing to generalize? I help companies diagnose and improve machine-learning systems that already have data, a baseline model, or a working prototype—but are not delivering the accuracy, robustness, or business value they should. I am a Kaggle Competitions Grandmaster with a peak global rank of #18, more than 10 years of applied data-science experience, 13 national and international competition awards, and over $350,000 in prize money. I have also led production machine-learning, forecasting, and optimization systems deployed across thousands of connected devices worldwide. HOW I CAN HELP • ML model audits and improvement Identify leakage, weak validation, poor feature choices, calibration issues, error patterns, and missed ensemble opportunities. • Predictive modeling Build high-performing models for tabular data, classification, regression, ranking, and other difficult prediction problems. • Forecasting and time series Develop or improve demand, load, operational, customer, revenue, and other forecasting systems. • Feature engineering, tuning, and ensembling Improve existing baselines through rigorous experimentation and competition-grade modeling methods. • Production ML pipelines Turn notebooks and prototypes into reliable, reproducible, maintainable workflows. • Applied AI and LLM systems Implement practical classification, extraction, evaluation, automation, and decision-support workflows with measurable value. WHY CLIENTS HIRE ME • Kaggle Competitions Grandmaster, peak global rank #18 • $350,000+ in competition prize money • 13 forecasting, prediction, and optimization awards • Former Chief Data Scientist at Shifted Energy • Previous senior data-science roles at H2O.ai and Iterable • Experience with enterprise AutoML, terabyte-scale prediction, forecasting, optimization, and production deployment • M.S. in Electrical Engineering and doctoral research in machine learning and optimization I am especially useful when you need an expert to quickly understand an existing system, determine what is limiting performance, and implement the highest-value improvements. Send me a brief description of your data, current model, evaluation metric, and desired outcome, and I will tell you where I see the strongest opportunities.
Earnings and projects
Last 6 months ÷ 2.7 months active (since Aug 2026)
Last 6 months: $6.1k
Billed in all: $6,490$6.5k
$6.1k in the last 6 months
Earnings, last 30 days
Skill rankings: Top 3% in Data Mining
Ranked by MRR among everyone on Upwork with the skill.
- Feature Engineering40 freelancers · 🇺🇸 #1Top 8%🇺🇸#1 of 4
- #5Time Series Forecasting110 freelancers · 🇺🇸 #1Top 5%🇺🇸#1 of 10
- #12Predictive Modeling125 freelancers · 🇺🇸 #4Top 10%🇺🇸#4 of 28
- #25Information Analysis523 freelancers · 🇺🇸 #6Top 5%🇺🇸#6 of 119
- #51Statistical Analysis987 freelancers · 🇺🇸 #15Top 5%🇺🇸#15 of 234
- #67Machine Learning Model1,169 freelancers · 🇺🇸 #13Top 6%🇺🇸#13 of 171
- #95PyTorch1,585 freelancers · 🇺🇸 #22Top 6%🇺🇸#22 of 160
- #170Deep Learning2,743 freelancers · 🇺🇸 #31Top 6%🇺🇸#31 of 221