# ArgonBI Systems Inc. — Full Context

> Canadian data intelligence company building environmental intelligence platforms, ML systems, and geospatial tools for British Columbia.

## Company Overview

ArgonBI Systems Inc. is a data intelligence company incorporated in British Columbia, Canada. We specialize in building production-grade software that transforms complex environmental, geospatial, and scientific data into accessible tools for decision-makers, researchers, and the public.

Our core competency is taking messy, fragmented government and open data sources and turning them into reliable, fast, production systems. We operate at the intersection of machine learning, geospatial analysis, and full-stack software engineering.

### Focus Areas
- Wildfire risk prediction and management
- Wild salmon conservation and monitoring
- Seismic hazard visualization
- Environmental and property due diligence research
- Government data intelligence

## Products & Platforms

### Infernis — Wildfire Risk Prediction Engine

**URL:** https://infernis.ca
**Source Code:** https://github.com/argonBIsystems/infernis
**Status:** Live, open-source (open-core model)
**License:** Open source

Infernis is an ML-powered wildfire risk prediction engine for British Columbia. It generates daily fire risk assessments at 1km resolution across the entire province.

**Technical Architecture:**
- Model: XGBoost gradient-boosted ensemble classifier
- Training data: 10 fire seasons (2015–2024) of BC wildfire history
- Data sources: 21 open datasets including:
  - BC Wildfire Service fire perimeters and weather stations
  - Environment Canada weather observations and forecasts
  - NASA FIRMS active fire detections (MODIS + VIIRS)
  - BC land cover and vegetation data
  - Topographic features (elevation, slope, aspect)
  - Historical fire frequency and burn severity
- Grid: 84,000+ cells at 5km resolution, 1km resolution for high-risk areas
- Features: 24–28 engineered features per grid cell
- Performance: 0.962 AUC-ROC (5km), 0.974 AUC-ROC (1km)
- Update frequency: Daily automated pipeline
- Stack: Python, FastAPI, XGBoost, PostGIS, Redis, Railway
- Frontend: Interactive risk map with temporal controls

**Why It Matters:**
British Columbia experiences devastating wildfire seasons. There is no other publicly accessible, ML-powered wildfire risk prediction tool for BC. Infernis fills a critical gap between raw weather data and actionable risk information for communities, emergency managers, and the public.

### SalmonWatch BC — Salmon Migration Monitoring

**Source Code:** https://github.com/argonBIsystems/salmonwatch
**Status:** Open-source

SalmonWatch aggregates and visualizes salmon migration data for British Columbia. It combines DFO (Fisheries and Oceans Canada) stock assessment data, environmental sensor readings, and hatchery reports into interactive maps and time-series dashboards.

**Technical Architecture:**
- Frontend: Next.js 16, Deck.gl (WebGL maps), uPlot (time-series charts)
- Backend: FastAPI, PostGIS
- Data sources: DFO NuSEDS, DFO Stock Assessment, environmental monitoring stations
- Visualization: Interactive migration maps, stock-level time series, environmental correlation views

**Why It Matters:**
BC wild salmon are in crisis. Data about salmon populations is fragmented across government databases, making it difficult for researchers, First Nations, and conservation groups to understand the full picture. SalmonWatch makes this data accessible and visual.

### Cascadia Explorer — Seismic Visualization

**Source Code:** https://github.com/argonBIsystems/cascadia-explorer
**Status:** Open-source

Interactive 3D visualization of the Cascadia Subduction Zone — the fault system that poses the greatest seismic risk to the Pacific Northwest.

**Technical Architecture:**
- 3D engine: CesiumJS
- Frontend: React, TypeScript
- Data: USGS real-time earthquake feed, slab geometry models, GPS plate motion vectors
- Features: Real-time seismicity at true depth, slab surface rendering, earthquake scenario modeling

## Technical Capabilities

### Machine Learning & Predictive Intelligence
- Production ML pipelines with XGBoost, PyTorch, LightGBM
- Ensemble model design and hyperparameter optimization
- Risk scoring and classification systems
- Time-series forecasting
- Feature engineering from geospatial and temporal data
- Model monitoring and automated retraining

### Full-Stack Software Engineering
- Next.js / React frontends with TypeScript
- FastAPI and Express.js REST API backends
- PostgreSQL and PostGIS databases
- Redis caching and session management
- Authentication, RBAC, multi-tenant architecture
- Comprehensive automated testing (unit, integration, e2e)
- CI/CD pipelines

### Geospatial Analysis
- PostGIS spatial queries and analysis
- Raster data processing (GeoTIFF, NetCDF)
- Real-time risk mapping at high resolution
- Deck.gl WebGL-accelerated map visualization
- Mapbox tile serving and styling
- CesiumJS 3D globe visualization
- Open data integration (government, satellite, sensor networks)

### Data Engineering
- ETL pipeline design and automation
- Government open data ingestion and normalization
- Data quality validation and monitoring
- API design and documentation
- Database schema design and optimization
- Scalable data processing workflows

## Team

### Arnold Dogelis — Founder & CEO
- Full-stack software engineer with deep ML expertise
- Director at Fraser Valley Salmon Society
- Built multiple production systems including Infernis, SalmonWatch, and Cascadia Explorer
- Based in British Columbia, Canada

## Infrastructure

- **Hosting:** Railway (primary), Vercel (static sites)
- **Databases:** PostgreSQL + PostGIS
- **Caching:** Redis
- **CI/CD:** GitHub Actions
- **Monitoring:** Custom health checks, PostHog analytics

## Open Source Commitment

All of our public-facing intelligence platforms are open source. We believe environmental and public safety data tools should be transparent and auditable. Our open-source work demonstrates our engineering approach — the architecture, testing practices, and code quality that we bring to every project.

## Contact

- **Website:** https://argonbi.com
- **GitHub:** https://github.com/argonBIsystems
- **Contact Form:** https://argonbi.com/#contact
- **Location:** British Columbia, Canada