Skip to content

DaScient/Energy

Repository files navigation

ARES-E: Agentic Resilience & Evaluation System for Essential-Infrastructure

The DaScient Integrated Infrastructure Suite (EWIS, WOIK, PHIAK)

ARES-E is a unified evaluation harness and toolkit designed to standardize telemetry, events, and operational metrics across the "Critical Quad" of national security: Data/Energy, Water, and Public Health. This suite provides the deterministic engines, plug-in architectures, and privacy-aware schemas necessary to evaluate AI agents and human-machine teams under real-world operational stress.

Live Dashboard

Launch Dashboard →

The analytics dashboard provides a single-pane-of-glass operational view across all three modules with:

  • Real-time streaming telemetry (synthetic data engine at 0.5 Hz)
  • 18 configurable threshold alert rules with deduplication
  • 8 interactive Chart.js visualizations (dual-axis, radar, area, bar)
  • LOE composite benchmark scorecard
  • 12-week score forecasting with 95% confidence intervals
  • DDIL simulation results summary
  • 100% client-side — zero backend, zero tracking, zero PII/PHI

Focus: Data Center Planning, Grid Operations, and AI Workload Efficiency.

EWIS is a plug-and-play Python toolkit for operators and researchers managing the intersection of high-performance compute (HPC) and energy markets.

Key Capabilities:

  • Grid & Market Intelligence: Standardizes payloads for grid stress signals, carbon intensity, and real-time energy pricing.
  • Data Center Optimization: Tools for capacity planning, PUE (Power Usage Effectiveness) assisted attribution, and cooling optimization.
  • AI Efficiency Metrics: Benchmarking "Energy per Token" and model workload efficiency to assess the environmental and operational cost of AI deployment.
  • Weather Integration: Open-source RSS and weather-driven intelligence to forecast impacts on load, pricing, and infrastructure reliability.

Features:

  • Extensible Plugin Framework: Support for Python entry points and local plugin discovery for "hot-swappable" industry adapters.
  • CLI Diagnostics: Command-line interface for rapid batch runs and system diagnostics.
  • Visualization Helpers: Plotly-first (interactive) and Matplotlib (fallback) support for analytics.

Focus: Water Treatment, Distribution, and Hydraulic Operational Metrics.

WOIK standardizes telemetry and events across municipal and industrial water infrastructure, providing a "digital twin" logic for agentic evaluation.

Key Capabilities:

  • Infrastructure Telemetry: Standardized schemas for treatment plants, distribution networks, lift stations, and storage tanks.
  • Operational Risk Assessment: Reference metrics for leak likelihood, water quality risk, and pump specific energy.
  • Energy-Water Nexus: Integrated accounting for carbon and energy consumption associated with water pumping and treatment.
  • Event Standardization: Normalizes disparate sensor data into a strict Pydantic payload schema.

Features:

  • Deterministic Engine: A local execution environment that runs plug-ins without external dependencies.
  • Interactive Local Dashboard: An air-gapped HTML/JS dashboard that visualizes report JSON for sensitive site operations.
  • Pydantic Schema Enforcement: Ensures data integrity across heterogeneous sensor networks.

Focus: Privacy-Aware Health Operations and Early Warning Systems.

PHIAK is a plug-in oriented toolkit for aggregated public health operations analytics, designed specifically to avoid the ingestion of PII/PHI.

Key Capabilities:

  • Capacity Signaling: Tracks ED beds, ICU occupancy, ventilator availability, and staffing/supply levels.
  • Incidence & Surveillance: Standardizes signals for cases, test positivity, and syndromic surveillance (wastewater, outbreak indicators).
  • Privacy Guardrails (Non-Negotiable):
  • Zero Individual Data: No identifiers, free-text notes, or address-level geolocation.
  • Aggregation by Design: All metrics are counts, rates, or rolling summaries.
  • Cell Suppression: Optional minimum cell count suppression to prevent re-identification in small populations.

Features:

  • Static Dashboard: Air-gapped HTML + JS + CSS dashboard—perfect for secure, JWICS-level environments.
  • Deterministic Engine: Provides reproducible early warning indices and report generation.
  • Documentation-First: Requires specific documentation for every metric and data source to ensure transparency and safety.

Shared Architecture & Use Cases

The Plugin Contract

All three kits share a "Plugin Contract," allowing different teams to integrate proprietary systems (SCADA, EHR, Data Center Management) without rewriting the core analytics engine.

Deployment & Interoperability

  • Notebook-ready: Designed for Data Scientists and Researchers to iterate in Jupyter/VS Code.
  • Air-Gapped Ready: All dashboards and reports are generated as static files, requiring zero internet connectivity for visualization.
  • Vendor Agnostic: Standardizes payload schemas so that any AI model or agent can be evaluated against these metrics via the ARES-E harness.

Contact & Teaming

For inquiries regarding DIU CSO teaming or implementation, contact: ARES-E@dascient.com


DaScient Energy, Weather, and Interoperability Suite

A notebook-ready toolkit for:

  • Deep and Generative AI analytics and open-source metrics design
  • Data center interoperability performance metrics and mitigation protocols under energy and grid stress
  • Open-source RSS and weather driven energy intelligence for data center planning

Included notebooks

  • notebooks/DaScient_DeepGenAI_Analytics_OpenSource_Metrics_Package.ipynb
  • notebooks/DaScient_DataCenter_Interop_EnergyCrisis_Notebook.ipynb
  • notebooks/DaScient_Weather_News_Energy_DataCenter_Planning.ipynb
  • ewis-toolkit/notebooks/01_grid_stress_eda.ipynb
  • ewis-toolkit/notebooks/02_genai_metrics.ipynb
  • public-health-infra-analytics-kit/notebooks/00_PHIAK_Tour.ipynb
  • water-ops-interop-kit/notebooks/00_WOIK_Tour.ipynb

Quickstart

python -m venv .venv
source .venv/bin/activate
python -m pip install -U pip
pip install -r requirements.txt
jupyter lab

Repo layout

  • notebooks/ - the main packages
  • src/dascient_suite/ - reusable modules (RSS, weather, energy proxies, reporting I/O)
  • docs/ - playbooks and glossary
  • reports/ - generated exports (optional)
  • data/sample/ - sample schemas

License

MIT - see LICENSE.

License

MIT License — see LICENSE for details.


DaScient, LLCSystematically addressing deep tech issues in critical infrastructure

GitHub · ARES-E@dascient.com


Disclaimer

Planning-grade analytics and operational scaffolding. Not a substitute for facility engineering, safety review, or regulatory compliance.

About

A vendor-agnostic evaluation harness for assessing AI agents across the "Critical Quad": Data Centers, Energy, Water, and Health. ARES-E integrates the EWIS, WOIK, and PHIAK deterministic engines to simulate mission-critical stress (DDIL conditions), evaluate human-machine teaming (HITL), & audit agentic tool-use through Pydantic-enforced schemas.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors