Industrial Engineer building practical AI decision systems for energy and industrial operations.
I work at the intersection of forecasting, optimization, financial modeling, and reliable software delivery—turning messy operational data into transparent, testable decisions.
LinkedIn: Connect with me
- Energy & industrial intelligence: forecasting, scenario analysis, constrained optimization, and operational analytics.
- Decision-support systems: explainable assumptions, downside-aware modeling, uncertainty analysis, and auditable outputs.
- Applied ML infrastructure: reproducible Python workflows, evaluation discipline, API delivery, and maintainable data products.
|
End-to-end energy decision support: time-series forecasting, constrained optimization, and Streamlit/FastAPI delivery. Signal: Forecasting · optimization · scenario analysis · operational analytics Python · pandas · scikit-learn · FastAPI · Streamlit |
Transparent real-estate investment screening with downside-first analysis, cash-flow modeling, debt constraints, and Monte Carlo risk. Signal: Cash-flow modeling · debt constraints · LP/GP waterfalls · IRR/NPV · Monte Carlo risk Python · Streamlit · Financial modeling · Underwriting |
|
Reproducible research exploring historical market behavior through technical, sentiment, and machine-learning signals. Signal: Feature engineering · XGBoost · LSTM · time-series experiments Python · Jupyter · pandas · NumPy · XGBoost |
Energy-market intelligence, oil & gas analytics, automation, and content workflows are also part of my current work. Signal: Proprietary data and active deployments stay private; public repositories show the engineering approach. Automation · APIs · analytics · operational workflows |
I am extending this work through focused contributions to Python, forecasting, scientific computing, energy systems, practical ML infrastructure, and trustworthy AI tooling. I prefer changes that are small enough to review and strong enough to keep:
- clear problem framing and explicit assumptions;
- reproducible tests and honest evaluation;
- maintainable APIs and documentation;
- regression coverage for edge cases and failure paths.
- GPT-OSS AB/CD parser fix: preventing malformed or ambiguous model outputs from becoming guessed evaluation answers.
I also contribute detailed, evidence-based reviews of open-source AI tooling, separating confirmed defects from security risks and improvement suggestions. Recent examples include reviews of shell allowlist enforcement, path-sandbox boundaries, and loopback-only development authentication.
What I work on
- Explainable forecasting and optimization for energy and industrial applications
- Scenario analysis, constrained planning, and uncertainty-aware decision support
- Process analytics and operational improvement
- Reproducible machine-learning workflows and usable data products
How I build
Understand the system. Make assumptions explicit. Build the simplest useful solution. Measure the result.
I value clear problem framing, trustworthy data, honest evaluation, and maintainable implementation. A model is only useful when it improves the decision around it.
I am open to thoughtful collaboration on applied AI, energy intelligence, forecasting, optimization, and transparent decision-support systems.
- Start a technical conversation: GridWise AI discussions and issues
- Review a decision-support system: AtlasRE
- Explore the full portfolio: all repositories
Please include a concrete use case, data boundary, metric, or reproducible example when opening a technical issue.