For investors & strategic partners

Atlas4D is a working deep-tech platform – not a slide deck. Built to scale from a single city to an entire country.

❓

The problem

Critical infrastructure, airspace and networks generate massive spatiotemporal data, but tools are siloed: radar, weather, IoT, video, events, network monitoring – all live in separate systems with no unified “4D brain”.

🧠

The Atlas4D answer

A unified 4D fabric (space + time) that ingests radar, weather, IoT, vision, events and network metrics into one model, with STSQL, NLQ and AI services on top – ready for real-time decisions.

📈

Status today

22+ microservices, live radar & weather fusion, threat forecasting, Network Guardian, event risk module with MALP calibration, full observability (Prometheus + Grafana) and running infrastructure.

🎯

Initial markets

Public safety & airspace security, telco & network operations, smart city command centers, large event operators and critical infrastructure providers.

🧱

Defensibility

Deep database + geospatial architecture, 4D data model, multi-sensor fusion engine, and domain-specific AI modules built on open standards (PostGIS, H3, Timescale, pgvector).

🤝

What’s next

1–2 lighthouse pilots with a city, telco or infrastructure partner, hardened multi-tenant cloud version, and regional rollout across SEE / EU.

Stage: pre-seed / seed, founder-led, bootstrapped.

Talk to the founder Open live demo

🔮 The Atlas4D Philosophy

"Atlas4D doesn't log events-it logs the history of movement."

Where is it?

Static data

Where has it been?

Temporal data

Where will it be?

Predictive analytics

⏱️

Temporality

We work with history, not just "state". No overwrites-versioned intervals (valid_from → valid_to).

🌍

Geospatial

STSQL: time + space + motion. Example: "Which drones approached the airport in the last 30 minutes?"

🚀

Forecasting

Trajectory analysis, vectors, and predicting positions N minutes ahead. Alerts before the event.

🏗️

Microservices

Dozens of specialized services-independent yet coordinated (IoT, AI, Weather, Vision).

🔐

Reliability

Idempotency, Snowflake ID, dual-layer auth. Resilient to restarts and network loss.

🤖

AI-Augmented

Digital twin of reality with LLM/NLQ, vector trajectories and predictive models.

🧠 What is Atlas4D, really?

Atlas4D is a 4D spatiotemporal data layer – a database-backed platform that mirrors the real world across space and time.

🗺️

4D Data Layer

Instead of separate systems for GIS, time-series, events and logs, Atlas4D keeps movement in space and time in one model – entities, trajectories, anomalies, threats and events.

🧩

Unified on Postgres

Built on PostgreSQL + PostGIS + TimescaleDB + H3 + pgvector, Atlas4D behaves like a 4D RDBMS: SQL-compatible, observable and production-ready, but with native support for geospatial, temporal and vector AI.

⚙️

From Sensors to Decisions

Radar, weather, IoT, video, network telemetry and events are ingested into a single spatiotemporal model – then exposed via maps, APIs, STSQL and NLQ for cities, telcos and infrastructure operators.

🚀

Built for Real-World Systems

Atlas4D is designed as an infrastructure layer: it runs on your hardware, next to your sensors and existing systems, and complements your current monitoring, GIS and analytics stack instead of replacing it overnight.

📊 STSQL - SQL for Motion

Our spatiotemporal extension for real-time queries

Spatiotemporal Query

Try →

-- Drones near the airport in the last 30 minutes
SELECT *
DURING ["now-30m","now"]
NEAR POINT(42.6977,23.3219)
WITHIN 5km
LIMIT 100;

Trajectory Forecasting

NLQ Query →

-- Predict position in 10 minutes and check a restricted zone
SELECT entity_id, predicted_position
FROM trajectories
WHERE PREDICT(+10m)
AND INTERSECTS(restricted_zone);

DURING [start,end] - time interval

NEAR POINT(lat,lon) - space

WITHIN radius - georadius

PREDICT(+time) - forecast

Key Capabilities

Everything for spatiotemporal data analytics

Geospatial Intelligence

PostGIS, H3 indexes and STSQL for massive real-time spatial data.

Real Time

TimescaleDB continuous aggregates and materializations for fast dashboards and alerts.

Visualization

MapLibre GL + MVT/PMTiles-millions of points, smooth render without blocking.

AI & NLQ

Integration with Ollama LLM: multilingual NLQ, STSQL generation, trajectory forecasts.

Microservices

IoT, Vision, Radar, Weather, Fusion… each service is a domain expert.

Anomalies & Threats

Anomaly detection and predictive alerts with fusion and ML.

Trust Quantification

✅ Shipped: Vision Trust (Wilson CI, split freshness, worst-label). Weather truthish partial. MALP γ=0.996.

👁️ Vision Trust NEW

Wilson CI 95% precision bounds, split freshness (recorder vs human loop), worst-label outliers, real-time UI badge.

Scenario Engine

✅ Shipped: Scenario contexts (Reality/WIND_STORM). 🧭 Roadmap: Monte Carlo what-if runner.

Quad-Temporal

✅ Shipped: valid/transaction in observations, issued_at in forecasts. 🟡 Expanding: as-of API across domains.

🏗️ Microservice Ecosystem

Specialized services working in sync

🌐 Frontend & Gateway

🎨

Frontend

MapLibre GL + React interface

🚪

API Gateway

Routing, auth, rate limiting

🔐

Auth Service

JWT authentication

📡 Data Ingestion

📥

Public API

External data ingestion

🔌

IoT Service

IoT device processing

📹

Cam Hook

Camera feed integration

🤖 AI & Analytics

💬

NLQ Service

Natural Language → STSQL

👁️

Vision Service

AI video/images

📊

Trajectory Embedding

Vectorization

🔮

Trajectory Prediction

Forecasts

⚠️ Anomaly Detection

🔍

Anomaly Service

Pattern-based

🧬

Anomaly Fusion

Correlation

⚡

Threat Forecaster

Predictive analysis

📡 Edge Processing

📡

Radar Edge

Realtime radar

🔊

Acoustic Edge

Audio analytics

🔄 Data Fusion

🎯

Fusion Coordinator

Orchestration

🔗

Fusion Service

Merging

🌤️ Environmental

🌡️

Weather Service

H3-indexed

☁️

Weather Fetcher

Integrations

⚙️ Execution & Legacy

🔧

WASM Exec

WebAssembly

🎰

WASM Predict

ML inference

🗄️

Monolith API

Legacy bridge

🔮 Predictive Threat Forecasting

AI-powered risk assessment combining LSTM trajectory predictions, weather hazards, and anomaly patterns

🧠

LSTM Predictions

86.9%

Trajectory risk score
20 trajectories analyzed

🌤️

Weather Hazards

22.5%

Wind & storm risk
Real-time conditions

⬡

H3 Resolution

~0.1 km² cells
174m edge length

⏱️

Forecast Window

30min

Prediction horizon
Configurable per request

Threat Level Classification

LOW

Score: 0.0 - 0.3

MEDIUM

Score: 0.3 - 0.7

HIGH

Score: 0.7 - 1.0

Multi-Source Fusion Algorithm

40% Weight

Anomaly Detection

Pattern-based risk scoring

35% Weight

Weather Hazards

Wind, storms, conditions

25% Weight

LSTM Trajectories

Prediction deviation analysis

API Endpoint: `POST /api/threats/forecast`

// Request
{
  "lat": 42.5048,
  "lon": 27.4626,
  "horizon_minutes": 30
}

// Response
{
  "threat_level": "medium",
  "threat_score": 0.313,
  "trajectory_score": 0.897,
  "weather_score": 0.255,
  "confidence": 1.0
}

Launch Threat Forecasting Interface →

🛡️ Network Guardian v1.0

Real-time network infrastructure monitoring with geospatial intelligence and semantic search capabilities

13,000+

Active Observations

H3 + PostGIS

Geospatial Indexing

pgvector

Semantic Search

90-Day

Retention Policy

📊

SNMP Monitoring

✓ Template-based metric collection
✓ Interface bandwidth tracking
✓ 30s ping / 5min SNMP cycles
✓ TimescaleDB hypertables

🗺️

Geospatial Intelligence

✓ H3 hexagonal indexing
✓ PostGIS geometry support
✓ MVT tile rendering
✓ MapLibre GL integration

🤖

Semantic Search

✓ Ollama embeddings (768-d)
✓ pgvector similarity search
✓ Hybrid semantic+geo+time
✓ Natural language queries

🚨

Alert Management

✓ Multi-channel notifications
✓ Email + webhook support
✓ Customizable thresholds
✓ Prometheus metrics export

⚡

STSQL Queries

✓ Spatiotemporal SQL adapter
✓ Complex correlation queries
✓ Cross-layer analysis
✓ Real-time aggregations

👁️

Unified Observations

✓ v_netguard_observations view
✓ Device + metric unification
✓ Geographic enrichment
✓ Temporal indexing

Architecture Overview

┌─ Network Devices
│   └─ SNMP/Ping Workers
│       ├─ ping_worker (30s intervals)
│       └─ snmp_worker (5min intervals)
│           └─ Template-based collection
│
├─ Data Layer
│   ├─ TimescaleDB (metrics hypertables)
│   ├─ PostGIS (geospatial geometries)
│   ├─ H3 (hexagonal indexing)
│   └─ pgvector (embeddings storage)
│
├─ Processing Layer
│   ├─ Ollama (embeddings generation)
│   ├─ Alert Worker (threshold monitoring)
│   └─ STSQL Adapter (query compilation)
│
└─ Presentation Layer
    ├─ MVT Tiles (MapLibre GL)
    ├─ REST API (FastAPI)
    ├─ Prometheus Metrics
    └─ Web Dashboard

Semantic Search Query: `GET /api/netguard/search`

// Query devices with performance issues near Burgas
{
  "query": "network devices with high latency in Burgas region",
  "location": { "lat": 42.5048, "lon": 27.4626 },
  "radius_km": 10,
  "time_range": "last_24h"
}

// Response with hybrid semantic + spatial + temporal results
{
  "results": [
    {
      "device_id": "lb.digicom.bg",
      "semantic_score": 0.89,
      "avg_latency_ms": 145.7,
      "distance_km": 2.3,
      "alert_count": 3
    }
  ]
}

Explore Network Guardian →

⚠️ Event Risk Assessment with MALP

AI-powered attendance prediction with Maximum Adjusted Linear Prediction (MALP) calibration for bias-corrected forecasts

γ = 0.996

MALP Concordance

-2.4%

Avg Bias Correction

BG + EN

Bilingual NLQ

Production

Ready System

🎯

MALP Calibration

✓ Maximum Adjusted Linear Prediction
✓ γ = 0.996 (excellent concordance)
✓ Automatic bias correction
✓ CCC monitoring & safety guards
✓ Transparent adjustments

🌐

Multi-Factor Analysis

✓ Weather impact (rain, temp)
✓ Health indices (flu, COVID)
✓ Calendar context (holidays)
✓ Traffic density (real-time)
✓ Historical patterns

⚖️

Risk Scoring

✓ Attendance risk (capacity)
✓ Weather risk (precipitation)
✓ Traffic risk (congestion)
✓ Combined risk score (0-1)
✓ Confidence intervals

📊

Full Transparency

✓ Base prediction (heuristic)
✓ Calibrated prediction (MALP)
✓ Adjustment delta (Δ)
✓ Calibration method exposed
✓ Confidence scores

🛡️

Safety Guards

✓ Minimum 10 events required
✓ CCC degradation detection
✓ Graceful fallback to heuristic
✓ Auto-calibration on startup
✓ Production logging

💬

NLQ Interface

✓ Bulgarian queries
✓ English queries
✓ Context-aware parsing
✓ Smart clarifications
✓ Detailed explanations

MALP Prediction Workflow

1. Base Prediction (Heuristic Model)
   └─ Multi-factor analysis → base_prediction: 4500

2. MALP Calibration (Bias Correction)
   ├─ γ coefficient: 0.9962
   ├─ Formula: calibrated = γ × base_prediction
   └─ Result: 4390 (adjustment: -110 or -2.4%)

3. Risk Assessment
   ├─ Capacity utilization: 4390 / 5000 = 88%
   ├─ Weather risk: precipitation probability
   ├─ Traffic risk: real-time observations
   └─ Combined risk score: 0.3 (Medium)

4. Response with Transparency
   ├─ predicted_attendance: 4390
   ├─ base_prediction: 4500
   ├─ calibration_method: "MALP"
   ├─ adjustment: -110
   └─ confidence: 0.65

Benefits:
✓ Reduced bias → more accurate predictions
✓ Conservative estimates → better planning
✓ Full transparency → trustworthy results

Event Organizers

Plan staffing, security, and resources with bias-corrected attendance forecasts

Venue Managers

Optimize capacity, pricing, and logistics with MALP-calibrated predictions

City Planners

Coordinate transport and emergency services with accurate crowd size estimates

Data Scientists

Access full transparency: base vs calibrated predictions with adjustment deltas

Risk Assessment API: `POST /api/events/risk/assess`

// Request - predict with MALP calibration
{
  "event_name": "Summer Music Festival",
  "event_type": "concert",
  "venue_capacity": 5000,
  "location": {
    "city": "Burgas",
    "lat": 42.5048,
    "lon": 27.4626
  },
  "date": "2025-07-15T20:00:00Z"
}

// Response - MALP transparency included
{
  "success": true,
  "predicted_attendance": 4390,
  "base_prediction": 4500,
  "calibration_method": "MALP",
  "adjustment_delta": -110,
  "adjustment_pct": -2.4,
  "risk_assessment": {
    "overall_risk": "Среден",
    "risk_score": 0.3,
    "attendance_risk": 0.88,
    "weather_risk": 0.15
  },
  "confidence": 0.65
}

MALP Statistics API: `GET /malp/stats`

// Response - calibrator status
{
  "calibrated": true,
  "n_events": 10,
  "gamma": 0.9961950734747499,
  "ccc_before": 0.9855,
  "ccc_after": 0.9733,
  "ccc_improvement_pct": -1.24,
  "data_source": "production",
  "last_calibration": "2025-11-11T14:20:22"
}

// Training endpoint
POST /malp/train?force=false

// Single prediction calibration
POST /malp/calibrate?prediction=7500

Technology Stack: NumPy (MALP calculations), asyncpg (database), FastAPI (API), Pydantic (validation), WeatherAPI, TimescaleDB, PostgreSQL views

Open Event Risk Map → View Event Risk Demo →

💼 For investors & strategic partners

Short version of the Atlas4D investor story: problem → what’s built → where we’re going.

1. The problem

Cities, telcos and infrastructure operators drown in spatiotemporal chaos: radar, weather, IoT, cameras, events, network monitoring – all in separate systems, with no unified 4D view and no "brain" that can predict what comes next.

Decisions are made on Excel, screenshots and phone calls, not on a live model of reality.

2. What’s already working

22+ microservices with 965k trajectory vectors, 25M queryable anomalies, live radar & weather fusion, threat forecasting, Network Guardian, event risk assessment with MALP
Real-time 4D map (H3 + PostGIS + TimescaleDB, MVT/PMTiles, up to ~225× faster rendering)
NLQ & STSQL interfaces for natural language spatiotemporal queries
Deployed, observable instance: atlas4d.tech / atlas4d.digicom.bg

3. Roadmap & stage

Atlas4D is founder-built, bootstrapped and deployment-proven. Next step: productizing the 4D data layer for 1–2 lighthouse customers.

We’re exploring a pre-seed / seed round to:

Run 1–2 full pilots (city / telco / infrastructure operator)
Harden the SaaS platform (multi-tenant, self-service onboarding)
Grow the open SDK ecosystem (Rust / Go / Python clients)

👉 Full investor brief: atlas4d.tech/investors-brief

🚀 Platform Updates

Recent development milestones from Sprint 1–13

💬

NLQ Conversation Memory

Context-aware follow-up queries. Ask "What's the weather in Burgas?" then "And in Sofia?" – it remembers.

Sprint 12 • 40ms cached responses

⚡

LLM Performance Boost

Redis intent cache, Ollama pre-warming, weather fast-path. From 21s to 70ms for common queries.

Sprint 12 • 300x faster

📱

Mobile-Ready Navigation

Unified header across all pages with hamburger menu. 7 navigation items: Dashboard, NLQ, Threats, STSQL, NetGuard, Events, Ops.

Sprint 13 • Full responsive

🔥

Fusion Engine 23,000x Faster

Optimized spatial scoring and threshold-aware processing. Real-time threat zone rendering at 0.1ms.

Sprint 7 • Production-ready

🛡️

Network Guardian v1.0

13k+ observations, SNMP/Ping workers, 11 ML models per device, MVT tiles, semantic search with pgvector.

Sprint 9-11 • Live monitoring

📅

Event Risk + MALP

AI attendance prediction with bias-corrected MALP calibration. Weather, traffic, and capacity risk scoring.

Sprint 10 • γ=0.996 calibrated

Explore the Platform →

Atlas4D Use Cases

From safety to smart cities and industry

🛡️

Safety

Drone incursions, anomalous trajectories, predictive alerts.

🏙️

Smart Cities

Airspace, traffic, coordinated incident response.

📡

Telecom

Load forecasts, coverage optimization.

🏭

Industry

Tracking, predictive maintenance, logistics.

🚁

Drone Operations

Routes, collision avoidance, optimal trajectories.

🔌

IoT Analytics

Massive streams, anomalies, smart notifications.

🌐

Infrastructure

Temporal monitoring and change analysis.

🌡️

Meteorology

H3 grid, Open-Meteo integrations and correlations.

🔥

Wildfire Detection

Early warning, spread prediction, evacuation planning.

🌾

Precision Agriculture

Yield forecasting, irrigation optimization, pest detection.

4D platform connecting cities, networks and critical events into an infinitely extensible layer for intelligent decisions

For investors & strategic partners

The problem

The Atlas4D answer

Status today

Initial markets

Defensibility

What’s next

🔮 The Atlas4D Philosophy

Where is it?

Where has it been?

Where will it be?

Temporality

Geospatial

Forecasting

Microservices

Reliability

AI-Augmented

🧠 What is Atlas4D, really?

4D Data Layer

Unified on Postgres

From Sensors to Decisions

Built for Real-World Systems

📊 STSQL - SQL for Motion

Key Capabilities

Geospatial Intelligence

Real Time

Visualization

AI & NLQ

Microservices

Anomalies & Threats

Trust Quantification

👁️ Vision Trust NEW

Scenario Engine

Quad-Temporal

See the platform in action

🚗 Fleet Tracking

🗺️ Interactive Map

⚙️ Advanced Controls

🌦️ Weather + Anomalies

🏗️ Microservice Ecosystem

🌐 Frontend & Gateway

📡 Data Ingestion

🤖 AI & Analytics

⚠️ Anomaly Detection

📡 Edge Processing

🔄 Data Fusion

🌤️ Environmental

⚙️ Execution & Legacy

🔮 Predictive Threat Forecasting

LSTM Predictions

Weather Hazards

H3 Resolution

Forecast Window

Threat Level Classification

Multi-Source Fusion Algorithm

API Endpoint: POST /api/threats/forecast

🛡️ Network Guardian v1.0

SNMP Monitoring

Geospatial Intelligence

Semantic Search

Alert Management

STSQL Queries

Unified Observations

Architecture Overview

Semantic Search Query: GET /api/netguard/search

⚠️ Event Risk Assessment with MALP

MALP Calibration

Multi-Factor Analysis

Risk Scoring

Full Transparency

Safety Guards

NLQ Interface

MALP Prediction Workflow

Event Organizers

Venue Managers

City Planners

Data Scientists

Risk Assessment API: POST /api/events/risk/assess

MALP Statistics API: GET /malp/stats

API Endpoint: `POST /api/threats/forecast`

Semantic Search Query: `GET /api/netguard/search`

Risk Assessment API: `POST /api/events/risk/assess`

MALP Statistics API: `GET /malp/stats`