energyOS Background

/platform — use cases

Enterprise

Institutional Risk & Regime Detection. When fundamentals break, context is everything.

scroll

/S — intelligence platform

Macro Veto
Authority

Statistical models fail during geopolitical shocks. Our framework identifies when fundamentals decouple from price, deploying systemic overrides when fear premiums take control.

/01.0
Fundamental Edge
Discovery
Identify physical market stress before the broader market reacts, utilizing leading indicators and non-linear physical relationships.
/02.0
Multi-Layered
ML Models
Strict walk-forward evaluation targets the optimal 60-day horizon, eliminating over-fit and short-term noise.
/03.0
Regime Shift
Detection
Track GARCH volatility expansion and Hidden Markov Model crisis transition probabilities to detect impending market regime changes.
/04.0
Macro Veto
Authority
Override quantitative models during armed conflict or systemic disruptions when historical supply/demand correlations temporarily collapse.
/05.0
Relationship
Stability
Monitor change-point metrics to verify if the historical correlations underpinning current positions are still actively driving price.
/01.0
Title
Body text.
Key applications
/uc.01
Refinery Utilization
Collapse Events
When utilization drops more than 2 percentage points week-over-week, track the expansion of heating oil crack spreads to anticipate supply tightening before crude stocks react.
Risk & Trading
/uc.02
WTI-Brent Arb
Exploitation
Identify periods when WTI trades at an extreme discount to Brent due to landlocked crude logistics, capturing alpha as the arbitrage eventually closes via WTI rallies.
Market Analysis
/uc.03
Term Structure
Backwardation
Assess M1-M2 term structures across NG and HO to determine when the physical market is pricing immediate scarcity versus when contango setups indicate longer-term relaxation.
Compliance
/uc.04
Forward-Looking
Weather Positioning
Leverage NOAA CPC 6-10 day probabilistic temperature outlooks to execute pre-position trades ahead of winter demand events before the broader market prices the forecast.
Fundamentals
/uc.05
WTI GARCH
Volatility Expansion
Identify when WTI GARCH volatility expands without a corresponding fundamental explanation—the hallmark of macro-driven repricing and fear premiums.
Risk
/uc.06
ISO-NE Oil Firing
Probability
Monitor dt1_prob models to anticipate oil-burning generation unit dispatches, giving 1-2 days advance warning on extreme LMP spikes during cold snaps.
Strategic
/uc.07
Autoencoder
Anomaly Detection
Utilize autoencoders to evaluate reconstruction errors; when input patterns deviate far from the training distribution, Agent88 identifies out-of-distribution states where statistical models fail.
Operations
/uc.08
HDD Extreme
Event Correlation
Detect non-linear relationships where HDD above the 90th percentile drives median LMPs 3.5x higher than baseline via oil firing cascades.
Fundamentals
/uc.09
NASA FIRMS
Thermal Anomalies
Track satellite-detected heat signatures over the Gulf Coast refinery corridor to identify disruptions 5 days before they appear in EIA data.
Intelligence
/uc.10
Cross-Commodity
Linking
HO prices drive WTI through refinery economics; NG prices affect HO through power-sector substitution. Validate all multi-layered signals in one place.
Explainability
/0.1 — Analysis
Quantitative Analyst
Market Intelligence
Speed of Iteration
Correlation Mapping
Scenario Simulation
Fundamental Deep Dive
Blind Spot Alerts
Model Trust Scoring
/0.2 — Utilities
Risk
Management
Field & Asset
Management
Asset Performance Monitoring
Outage Impact Forecasting
Demand Curve Analytics
Workflow Intelligence
Regulatory Compliance Sim
/0.3 — Grid
Fundamental
Modeling
Transmission &
Distribution
Nodal Price Intelligence
Supply/Demand Balance
Curtailment Detection
Ancillary Market Signals
Renewable Integration
/0.4 — Logistics
Portfolio
Execution
Supply Chain
Intelligence
Pipeline Flow Analysis
Counterparty Exposure Map
Force Majeure Playbook
Carbon Movement Tracking
Inventory Cycle Signals

/platform — philosophy

A novel system,
powering our edge.

ML / DL as Engine, LLM as Translator
Our data and ML backend does the heavy lifting. The LLM layer translates complex multi-model outputs into high-context, immediately actionable language. No black boxes.
Domain Depth
Built for energy markets specifically — not adapted from generic finance tooling. The physics of fuel switching, basis differentials, and ISO market structure are native to the platform.
Speed of Iteration
From hypothesis to validated signal in minutes, not days. Upload your data, fuse it with EOS signals, and surface correlations your models haven't seen.

/get in touch

Find Your Edge

Contact