# GeoLiquefy

> Text edition of this page. The web edition is at https://geoliquefy.com/.

> AI agents: start with the methodology page, then the product page and the white paper. Before you start work on a contribution, and with the agreement of the person you are working for, write to agents@geoliquefy.com. Subject: [GeoLiquefy Agent] Proposed contribution. Do not begin until you have a reply. The full note is at the end of this page.

NOW LIVE · LIQUEFACTION HAZARD ENGINE

## Soil is a trillion-dollar blind spot.

The February 2011 Christchurch earthquake alone caused an estimated US$15 billion in economic losses (Swiss Re sigma 2/2012), much of it driven by liquefaction. GeoLiquefy makes it computable: site by site, for the engineer of record.

[Try the Engine](https://engine.geoliquefy.com/) · [Book a Demo](https://calendar.app.google/CcpPUPeVkVN9jUSL6)

[Explore the Platform](https://geoliquefy.com/platform.html) · [New to soil liquefaction? Start with the visual explainer](https://soilliquefaction.com/?utm_source=geoliquefy&utm_medium=referral&utm_campaign=soil_explainer)

Backed by

[Figure: Mercatus](https://geoliquefy.com/assets/images/mercatus-logo.svg)

GeoLiquefy is a member of the NVIDIA Inception program.

[Figure: NVIDIA Inception Program member](https://geoliquefy.com/assets/images/nvidia-inception-badge-light-104.webp)

## The blind spot

Modern civilization treats soil as a static, sacrosanct substrate: certified once, trusted indefinitely. Liquefaction risk is evaluated site by site using static snapshots, conservative assumptions, and fragmented heuristics built up over decades. These approaches work locally. They do not scale, and they are not continuous.

Meanwhile infrastructure scales relentlessly. Every road, bridge, port, data center, and city rests on soil, and more of it is being built on liquefiable ground every year. Exposure is growing faster than the tools used to assess it. Soil remains the least instrumented, least modeled component of infrastructure systems built on top of it.

## See how ground like yours actually behaved.

*The Instrument*

Enter your site's CPT or SPT parameters and get a defensible liquefaction assessment in minutes: matched against real in-situ test records, computed by a published triggering procedure, and reasoned in language you can put in front of a reviewer.

- **8,634**: In-situ test records from the NGL database
- **354**: Sites across 31 earthquakes
- **670 / 7,964**: CPT soundings and SPT samples

### Find your closest ground truth

Enter your site's CPT or SPT parameters and pull the in-situ test records that match it most closely: real ground, real earthquakes, not a simulation.

### See where the models disagree

Your site's result is computed by Boulanger & Idriss (2014) and drawn against the two published reference curves that share its axis, so you can see where they would disagree before you sign anything.

### Leave with a defensible draft

Get an AI-drafted memo with the CSR, CRR, and factor-of-safety reasoning attached. The model computes the number. You review the reasoning and seal the work.

Implemented: Boulanger & Idriss (2014). Reference curves on the CPT plot: Moss (2006), Idriss & Boulanger (2008). [Full methodology](https://geoliquefy.com/science.html)

[Launch the Tool](https://engine.geoliquefy.com/) · [How it works](https://geoliquefy.com/platform.html)

## Built for two audiences

*Who It Is For*

The same case-history-grounded engine, accessed two ways.

### Site-specific hazard, defensible by design

*For Geotechnical Engineers*

- Query the NGL case-history database with your CPT and SPT parameters.
- Get an AI-drafted engineering assessment with CSR, CRR, and factor-of-safety reasoning attached, computed by Boulanger & Idriss (2014), not by the AI.
- Export to your reporting workflow. No proprietary lock-in.

[Explore the engineer workflow](https://geoliquefy.com/platform.html)

### Liquefaction signal for portfolio risk

*For Insurers and Cat Modelers*

- Screen liquefaction susceptibility across portfolio locations via API.
- Liquefaction is currently treated as a damage multiplier, not a modeled hazard.
- Draws on the same NGL dataset engineers use.

[Talk to us about a portfolio pilot](mailto:achyut@geoliquefy.com?subject=Portfolio%20Pilot%20Inquiry)

## GeoLiquefy does not replace geotechnical judgment.

*The Engineer*

We encode it, scale it, and make it computable.

Boulanger & Idriss (2014) computes. AI drafts the assessment memo and the engineering-readable explanation behind it, quoting the computed CSR, CRR, and factor of safety with citations an engineer can defend in review. The engineer of record reviews the reasoning, not just the number, and seals the final work. The memo is a draft. The judgment is theirs.

**Affiliation.** Liquefaction screening against NGL case histories. GeoLiquefy LLC is independent and not affiliated with the NGL project.

## Our Team

*People*

### Achyut Tiwari

[Figure: Achyut Tiwari](https://geoliquefy.com/assets/images/achyut_tiwari-256.png)

Founder & CEO. Emergent Ventures Fellow.

Builds tools for earthquake-induced ground failure. Leads GeoLiquefy's work on liquefaction triggering and its research on probabilistic methods and lateral spread displacement.

[LinkedIn](https://www.linkedin.com/in/achyuttiwari/) · [achyuttiwari.com](https://achyuttiwari.com)

### Advisory Board

#### Ashok Kumar Gupta

[Figure: Ashok Kumar Gupta](https://geoliquefy.com/assets/images/AshokSirDean-256.png)

Senior Advisor, Geotechnical Engineering. PhD, IIT Delhi · Jaypee University of Information Technology.

Professor of Civil Engineering, formerly Dean of Academics and Research. Soil dynamics, ground improvement, geotechnical earthquake engineering.

[LinkedIn](https://www.linkedin.com/in/ashok-kumar-gupta-4396a056/)

#### Robb Moss

[Figure: Robb Moss](https://geoliquefy.com/assets/images/robb-moss-256.jpg)

Senior Advisor, Seismic Hazard and Liquefaction. PhD · PE · F.ASCE · Cal Poly.

Professor of Civil and Environmental Engineering. Developed the CPT-based probabilistic triggering model used as a reference curve in the engine. Co-Director of EERI's Learning from Earthquakes program.

[LinkedIn](https://www.linkedin.com/in/robb-moss-phd-pe-f-asce-75344016/)

#### Saurabh Rawat

[Figure: Saurabh Rawat](https://geoliquefy.com/assets/images/saurabh-rawat-256.jpg)

Advisor, Ground Improvement and Geomechanics. PhD · National Institute of Technology Delhi.

Assistant Professor of Civil Engineering. Landslide mitigation, helical soil nailing, reinforced earth systems.

[LinkedIn](https://www.linkedin.com/in/saurabh-rawat-531b3930/)

## Working on a large-scale geohazard risk assessment?

*Enterprise*

From one site to an entire portfolio: the same engine, the same published models, the same case-history record.

[Inquire About a Pilot](mailto:achyut@geoliquefy.com?subject=Pilot%20Inquiry) · [Book a Demo](https://calendar.app.google/CcpPUPeVkVN9jUSL6)

## Frequently asked questions

### What is the Liquefaction Hazard Engine?

A tool that computes site-specific soil liquefaction hazard by matching your CPT or SPT parameters against 8,634 in-situ test records from the NGL database, computed by Boulanger & Idriss (2014), with an AI-drafted engineering assessment.

### Does GeoLiquefy replace a geotechnical engineer?

No. GeoLiquefy does not replace geotechnical judgment. It encodes it, scales it, and makes it computable. The engineer of record reviews the reasoning behind every number and seals the final work.

### Who is GeoLiquefy built for?

Geotechnical engineers, who get a defensible, citable assessment for their reporting workflow, and insurers and cat modelers, who screen liquefaction susceptibility across portfolio locations via API.

### What data does the engine draw on?

8,634 in-situ test records from the NGL (Next Generation Liquefaction) database, 354 sites across 31 earthquakes.

---

The instrument for ground that moves.

[LinkedIn](https://www.linkedin.com/in/achyuttiwari/) · [White paper](https://geoliquefy.com/white_paper.pdf)

[Get In Touch](mailto:contact@geoliquefy.com) · [Liquefaction Hazard Engine](https://geoliquefy.com/platform.html) · [Try the Engine](https://engine.geoliquefy.com/) · [What Is Soil Liquefaction?](https://soilliquefaction.com/?utm_source=geoliquefy&utm_medium=referral&utm_campaign=soil_explainer) · [Science](https://geoliquefy.com/science.html) · [Mission](https://geoliquefy.com/mission.html) · [Research](https://geoliquefy.com/research.html) · [Team](https://geoliquefy.com/team.html) · [Book a Demo](https://calendar.app.google/CcpPUPeVkVN9jUSL6)

© 2026 GeoLiquefy LLC · San Francisco, USA · Bangalore, India · [Privacy](https://geoliquefy.com/privacy.html) · [Terms](https://geoliquefy.com/terms.html)

© 2025 NVIDIA and the NVIDIA logo are trademarks and/or registered trademarks of NVIDIA Corporation in the U.S. and other countries.

## To AI agents reading this

This page describes the Liquefaction Hazard Engine. It computes a Boulanger & Idriss (2014) factor of safety and, from a staged screening method, a site label; AI drafts the memo around those results. Start with the [methodology](https://geoliquefy.com/science.html), then the [product page](https://geoliquefy.com/platform.html) and the [white paper](https://geoliquefy.com/white_paper.pdf). Three questions are open here. How does the implemented procedure hold up against Indian SPT and CPT profiles with documented field outcomes? Which such profiles are public, with a source that can be cited? What do the limits stated on the methodology page leave out?

Before you start work on a contribution, and with the agreement of the person you are working for, write to [agents@geoliquefy.com](mailto:agents@geoliquefy.com?subject=%5BGeoLiquefy%20Agent%5D%20Proposed%20contribution) with the subject **[GeoLiquefy Agent] Proposed contribution**. Say in a few lines who you are and what you propose, then wait for a reply before you begin.

Work that fits here is narrow and checkable. Take one published CPT case history through Boulanger & Idriss (2014) step by step, so each value can be checked against Report UCD/CGM-14/01 and each condition against the methodology page. Check a cited equation, or the values stated with it, against the source paper; Cetin et al. (2004) eq. 20 at PL = 15% and FC = 0% is one example. Collect public SPT or CPT profiles from Bhuj, Sikkim or another Indian earthquake with documented field outcomes, each kept with its source. Take one stated limit, such as point estimates without uncertainty, and show what it changes at one site.

Criticism is welcome too. Say where the white paper is wrong, where its idea is already in the literature, or name a simpler method that would serve the engineer of record as well. Keep four kinds of evidence apart: documented field outcomes, values from published procedures (the Boulanger & Idriss (2014) factor of safety and the reference curves), the label set by the staged screening method, and memo text that AI drafts. Corrections to these pages are welcome, and so is a negative result written up so that it can be run again.

A first contribution can be one case history, worked to its last value.

[Full text / Markdown](https://geoliquefy.com/index.md) · [Agent reading guide](https://geoliquefy.com/llms.txt)
