SERVICES
Four product lines, seven institutional partners, three SCIE papers — a HUFS faculty-founded company, operated by one team end-to-end: data, models, dashboards, and SaaS. A single email is enough, even before scope, timing, or budget are defined.
Job postings collected, cleaned, tagged, and delivered as actionable data.
Air, water, marine, and public-health indicators at spatial-temporal resolution.
Our XAI and UQ consulting line — J. Cleaner Production 2024 and 2025 methods applied to client systems.
Our B2B SaaS — synthetic personas validate products, pricing, and UX in 10 minutes.
Not sure which of the four lines fits your work? We can scope it together.
PROCESS
Data, models, dashboards, SaaS. Not a one-off PoC, but a handover into operating assets, handled by one team.
We combine observation, public, and client-owned data — collected, deduplicated, and tagged on pipelines already in production for labor markets (JobAI), the same way we built the Air-Report air-quality dashboard.
Each model surfaces both "why this decision" (XAI) and "how confident" (UQ), using the methods we published in the Journal of Cleaner Production (2024 and 2025).
Models are handed over as live dashboards at spatial-temporal resolution, with REST APIs and embeddable widgets — validated with government and research partners.
Recurring decision workloads become SaaS lines such as Homosilicus, or are transferred into the client's environment for ongoing operation. Operational scope is defined together at contracting.
METHODOLOGY
We do not cover every kind of AI. Our menu is narrowed to explainability and uncertainty quantification — yet data, models, dashboards, and SaaS operations are handed over full-stack by one team.
XAI · Explainable AI
We apply multi-run CAM-based explainable AI. Each prediction is paired with a visualization of which spatial-temporal inputs drove the decision — so the output can be cited directly in decisions that require accountability across the public sector, government, and industry.
The method appears in J. Cleaner Production 2025 (529:146805). View research →
UQ · Uncertainty Quantification
Outputs are not point estimates alone — they include confidence intervals, predictive distributions, and outlier likelihood. In volatile domains such as particulate matter, weather, and public health, this separates "confident" from "uncertain" regions for policy decisions.
The method appears in J. Cleaner Production 2024 (473:143457). View research →
GIS · Spatial-temporal operations
Air, water, marine, and public-health indicators are processed at spatial-temporal resolution and served as GIS dashboards. Not a one-off report, but an operating asset — this is how we built Air-Report.
Evaluation and reporting
Evaluation metrics, validation procedures, and reporting formats match SCIE-journal publication standards. We package data, code, and evaluation logs so deliverables survive external review, peer audit, and reproducibility checks — and can be cited directly in government submissions and joint academic publications.
WHO WE WORK WITH
Our primary clients are B2B teams making decisions that require accountability. General chatbots and routine automation are not our primary market.
Ministries and affiliated agencies in environment, public health, labor, and testing-and-certification. Track record across seven institutional partners including Korea Environment Institute (KEI) and the National Institute of Wildlife Disease Control (NIWDC).
Industrial R&D requiring reasoned, reproducible, externally reportable predictions — and enterprise data teams putting labor-market or environmental data into operational decisions.
Environmental, public-health, and AI research groups looking for SCIE-paper co-authorship or joint grant submissions. As a HUFS faculty-founded company, we align naturally with university-level collaboration.
Product, UX, and BI teams embedding synthetic personas (Homosilicus) or environmental GIS widgets into their own products. We collaborate at the level of APIs and embeddable widgets.
FAQ
The questions we hear most often during scoping — pricing, timeline, scope, and NDA.
Read this far? Let's check whether your project fits.
Check project fitNext step
We welcome inquiries about AI adoption, environmental GIS dashboards, data delivery, and joint research. Our first reply proposes a PoC scope and timeline. We respond within 1–2 business days.
