A geospatial AI platform, engineered in Japan. Built and ready — worldwide.
11 productized verticals on a single high-performance pipeline. NL Query, agent orchestration. Built end-to-end — from raw GPS to interactive maps.
If any of this sounds familiar:
- Your pipeline buckles at TB-scale, or takes overnight to refresh.
- Your team needs to write SQL just to ask basic questions of geospatial data.
- AI assistants hallucinate over your geo data instead of actually querying it.
Ask in plain language. The map answers.
No SQL, no filter-cage UI. Type a question the way you'd say it out loud — the system reads your warehouse schema and returns the answer on the map.
“Show the links with less than 50 km/h speed within 3 km of gas stations” — plain English in; the matching Tokyo links on the map and the SQL it generated, out.
Both the Tokyo and Osaka datasets were generated by traffic microsimulation — roughly one million simulated vehicle trips per city, about two million combined.
See capabilities →How I work
Four packages, fixed scope, indicative pricing.
Foundation
Typical project: ¥4M+Production-ready ingestion + calibration pipeline on the data you already collect.
A real-estate agency whose listing-feed updates from a dozen partners arrive in a dozen incompatible schemas — schema-normalized ingestion with calibration on a real day's traffic, queryable as one source.
- Schema + columnar storage layout
- Streaming or batch ingestion with backpressure + retries
- Calibration sprint on a sample dataset
- Observability: per-stage timings + correctness checks
Visualization & Query
Typical project: ¥6M+TB-scale data made browsable: tile servers + NL query over your warehouse.
A property listing platform where customers fight a filter-cage UI (price + station + size + commute) to express fuzzy preferences like 'family-friendly, near a green line, 30 minutes from Shibuya, under ¥200K.' NL query over the listing data accepts the question as written.
- High-performance vector tile server over your columnar storage
- Modern browser-based geospatial visualization
- Natural-language query layer over your warehouse
- Latency budget + caching strategy
Agentic Layer
Typical project: ¥8M+Domain-specific agents and the orchestration substrate they run on.
An energy site-selection workflow where agents auto-investigate each candidate lot — 'is it grid-connectable, is the substation at capacity, what is EV charging demand within 2 km' — with citations to the source data instead of analyst hours per site.
- Tool server exposing your domain capabilities
- Agent orchestration (structured tool I/O, retries, traces)
- Evaluation framework for agent behaviors
- Cost + latency telemetry
Microsimulation
Typical project: ¥3M+Vehicle-level traffic simulation for scenario analysis, synthetic data, and what-if planning.
A property developer modeling 'what does opening this new shopping center do to neighborhood traffic' before construction commits the capital, with side-by-side metric comparison across alternative site plans.
- Road network build for your geography (OSM-based)
- Demand calibration against your counts, OD data, or sample probes
- Baseline simulation + N alternative scenarios
- Synthetic trajectory output for downstream analytics
Project doesn't fit one of these shapes? Custom engagements are also available — get in touch to scope yours.
Selected evidence
TB-scale GPS data map-matched on a single commodity server. A pilot cohort of 10M+ records completes in single-digit minutes — substantially faster than common alternatives on the same hardware.
Methodology →Application domains
Government data programs · Automotive R&D through universities · Public-sector research
About
High-performance geospatial AI platforms targeting government, automaker, and research-domain applications — in Japan and internationally. 11 productized verticals on a single pipeline — from AADT volume estimation to bridge inspection. Technical brand for a Japan-based engineer specializing in TB-scale geospatial systems and agentic AI.
More →Let's talk
Send a message. Reply within one business day.