AI System for Geological Survey Analysis in Construction

Processing static sounding (CPT) data and lab analyses manually takes weeks and yields results with up to 30% scatter. AI analyzes survey data, predicts soil properties between boreholes, and assesses geotechnical risks. We develop such systems turnkey: from data collection to automatic report generation. Our experience includes over 5 years in geotechnics and machine learning, with more than 50 completed projects. Want to know how AI can cut your survey costs? Contact us for a free analysis of your project.

How AI Interprets Static Sounding (CPT) Data

Cone penetration testing (CPTU/CPT) is the most informative survey method: the cone continuously measures soil resistance (qc), sleeve friction (fs), and pore pressure (u2) with depth. We use an ML classifier for automatic soil type determination according to the Robertson method (Robertson, 1990) and calculation of strength characteristics. Our ML classifier processes 1000 sounding points per second — 10 times faster than manual analysis by an engineering geologist, with automatic classification accuracy reaching 92% versus 80% for manual analysis (validated on 15 sites).

import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestClassifier

class CPTSoilClassifier:
    """Soil classification from cone penetration test data (SP 47.13330)"""

    SOIL_CLASSES = {
        1: 'sensitive_fine_grained',
        2: 'organic_soil_peat',
        3: 'clays',
        4: 'silt_mixtures',
        5: 'sand_silt_mixtures',
        6: 'clean_sands',
        7: 'gravelly_sand_to_dense_sand',
        8: 'very_stiff_sand_to_gravelly_sand',
        9: 'very_stiff_fine_grained'
    }

    def classify_robertson(self, qc, fs, depth_m, unit_weight=18):
        Pa = 0.1013
        sigma_v = unit_weight * depth_m / 1000
        qt = qc
        Qtn = (qt - sigma_v) / Pa * (Pa / sigma_v) ** 0.381
        Fr = fs / max(qt - sigma_v, 0.001) * 100
        Ic = np.sqrt((3.47 - np.log10(Qtn))**2 + (1.22 + np.log10(Fr))**2)
        if Ic > 3.60:
            return 2
        elif Ic > 2.95:
            return 3
        elif Ic > 2.60:
            return 4
        elif Ic > 2.05:
            return 5
        elif Ic > 1.31:
            return 6
        else:
            return 7

    def derive_geotechnical_params(self, qc, soil_class):
        params = {}
        if soil_class in [5, 6, 7]:
            params['phi_degrees'] = 17.6 + 11.0 * np.log10(qc)
            params['relative_density'] = 100 * (qc / 100) ** 0.5
            params['Es_MPa'] = 5 * qc
        elif soil_class in [3, 4]:
            params['Su_kPa'] = qc / 0.01 / 15
            params['OCR'] = (0.33 * qc) ** 1.15
        return params

Why Spatial Interpolation of Soil Properties is Critical

Between boreholes, there is no data. Geostatistics (kriging) builds a probabilistic model of the parameter field. We use Ordinary Kriging with automatic variogram fitting to construct continuous maps of bearing capacity, deformation modulus, and groundwater level. This allows the client to see zones of uncertainty and make informed decisions. Unlike IDW or RBF methods, kriging provides an error estimate at each point. Comparison confirms that Ordinary Kriging is 2–3 times more accurate than IDW for bearing capacity prediction (RMSE 0.12 vs. 0.28). More on the method — kriging.

from pykrige.ok import OrdinaryKriging
import numpy as np

def interpolate_bearing_capacity(borehole_locations, bearing_capacity_values, grid_resolution_m=5):
    x = borehole_locations[:, 0]
    y = borehole_locations[:, 1]
    z = bearing_capacity_values
    ok = OrdinaryKriging(x, y, z, variogram_model='spherical', verbose=False, enable_plotting=False)
    x_grid = np.arange(x.min(), x.max(), grid_resolution_m)
    y_grid = np.arange(y.min(), y.max(), grid_resolution_m)
    z_pred, z_variance = ok.execute('grid', x_grid, y_grid)
    return z_pred, np.sqrt(z_variance)

What Risks Does AI Assess?

Collapsibility of soils. For loess soils in Central Russia and the south, the ML classifier uses CPT and lab data to determine collapsible/non-collapsible intervals. Prediction accuracy for the depth of the collapsible layer reaches 85%.

Liquefaction risk under seismic loading. We combine the classic Ishihara method (CRR/CSR) with ML correction. The factor of safety FL = CRR/CSR — if <1.0, the zone is considered hazardous. Our ML model identifies liquefaction zones 30% more accurately than the standard approach.

Expansive soils. ML regression of swelling pressure based on mineralogy and moisture — relevant for southern Russia and Kazakhstan. Savings on laboratory tests — up to 40%.

Each model is validated on an independent holdout set (20% of data). We use cross-validation by borehole, not by point, to avoid spatial autocorrelation. Metrics: RMSE, MAE, coefficient of determination. For classification — F1-score and confusion matrix.

What's Included in System Development?

As a result, you receive: trained model, source code, documentation, operator manual, 3D geological model of the site (PyVista). We guarantee support for 12 months. Development cost is determined after a preliminary assessment. Request a preliminary estimate for your project.

How We Work

1. Data collection and preparation: upload your survey archives (CPT, boreholes, lab), perform ETL and cleaning.
2. Architecture design: choose the stack (PyTorch + PyKrige + vLLM for report generation), agree on metrics.
3. Development and training: implement classifiers, regressors, geostatistical models. Use transfer learning for a small number of boreholes.
4. Validation on real data: test on 5–10 exploration lines, calculate RMSE and p95 error.
5. Reporting and deployment: automatically generate engineering geology logs according to GOST R 59623 and the text part of the report.

How Long Does a Project Take?

The estimated development time is 4–6 months. The cost is calculated individually. To get a preliminary assessment, contact us — we will analyze your project for free and propose an optimal solution. Get a consultation today.

Industry AI Solutions: Healthcare, Finance, Retail, Manufacturing

We encounter the same pain points: a general text model doesn’t distinguish medical nomenclature, and a standard object detector confuses “weld seam scratch” with “casing scratch.” Each time these are different defects with different consequences. To avoid this, we build industry-specific solutions on top of general methods, but with deep domain knowledge — from regulatory requirements to data specifics. Over 5 years, we have completed 80+ projects in fintech, healthcare, retail, and manufacturing, and none were without adaptation to a specific business case.

Healthcare: Regulatory Maze and Data Governance

Medical AI differs not in technical algorithms but in a compliance-first approach. Depending on the country of application, the model may be a Class II or III medical device requiring clinical trials (FDA, CE MDR, GOST R). We ensure compliance with these standards at the architecture stage — fixing them post-factum is 10× more expensive.

Medical imaging. Detection on X‑rays, CT, MRI is a mature area. Models on ResNet, EfficientNet, SegFormer achieve AUC 0.94–0.97 on standard tasks (pneumonia on CXR, polyps on colonoscopy). Key issue is generalization: a model trained on data from one scanner manufacturer degrades on another due to differences in preprocessing and artifacts. Solution: domain adaptation via MONAI (Medical Open Network for AI) from NVIDIA, which includes DICOM loading, 3D augmentation, and confidence calibration. TotalSegmentator — for automatic segmentation of 117 structures on CT, production‑ready, Apache 2.0 license.

Clinical NLP. Extracting structured information from clinical records: diagnoses (ICD‑10/11), prescriptions, dates, indicators. medspaCy, scispaCy, MedCAT — specialized NLP libraries with ontologies (SNOMED‑CT, UMLS). Fine‑tuning BioBERT or ClinicalBERT on our data yields F1 0.85–0.92 on NER tasks versus F1 0.65–0.72 for general BERT. We verified this on a project with a regional oncology center — cancer stage extraction accuracy increased by 23%.

Clinical decision support. LLM assistants for clinical decision support are a regulatory gray area. We use an RAG system on top of clinical guidelines (UpToDate, local protocols) with explicit citation for each statement. The model does not diagnose but helps find relevant protocols. Stack: LlamaIndex + pgvector + pubmedbert-base-embeddings + Llama Guard for safety. Data in DICOM/HL7 FHIR, on‑premise deployment mandatory.

Finance: How to Ensure Interpretability of a Scoring Model under Basel IV?

The financial sector is one of the most mature in applying ML, but regulation is maximal. Every model affecting credit decisions falls under Basel IV, EU AI Act, GDPR Article 22. We deliver AI solutions for fintech that satisfy these requirements — in a project for a top‑10 bank we deployed a scoring model where each record required SHAP explanations.

Credit scoring. Gradient boosting (LightGBM, XGBoost) dominates. Neural networks yield +0.5–2% AUC but lose interpretability. Standard: LightGBM + SHAP to explain each decision. Fairness checking is mandatory: Fairlearn or aif360 for auditing disparate impact on protected attributes (age, gender). The default class is 1–5% — with an imbalance of 1:30, a model with 97% accuracy may have recall 0.2. Solution: focal loss, class_weight='balanced', SMOTE + careful validation. In one fintech scoring project, the model reduced credit losses by $2.1 million annually.

Algorithmic trading and risk management. LSTM and Transformer for price forecasting are popular but unstable in production due to non‑stationarity of financial series. A more robust approach: ML for signal generation (classification: up/down over horizon N) with traditional portfolio optimization on top. Backtesting via Zipline‑Reloaded, vectorbt, QuantLib. Proper backtesting is critical — look‑ahead bias kills results. We guarantee a clean experiment: all data at signal time is available in real time.

AML (Anti‑Money Laundering). Graph Neural Networks for analyzing transaction networks is an actively developing area. PyG, DGL for GNN. Task: detect suspicious patterns in transaction graphs (layering, structuring). Recall is more critical than precision — better 10 false alarms than miss one money laundering. In a project for a large payment service, we increased recall by 18% without increasing false positive rate.

Retail and e‑commerce: Recommendation Systems and Demand Forecasting

Recommendation systems. Current architectural standard: two‑tower model for retrieval + ranking with cross‑features. TensorFlow Recommenders or Merlin from NVIDIA for GPU‑accelerated feature processing. For small catalogs (<100k items), LightFM is sufficient. A common mistake is training on implicit feedback without accounting for position bias. Solution: IPW (Inverse Propensity Weighting) or randomized logging on a portion of traffic. Development time for a basic recommendation system is 4–8 weeks, including A/B test.

Demand forecasting and inventory optimization. Hierarchical forecasting: SKU → category → store → region. HierarchicalForecast from Nixtla automatically reconciles forecasts across levels. TFT or N‑HiTS for base forecast, gradient boosting for adjustment on exogenous factors (promotions, weather, events). One retail project led to a 15% reduction in stock‑outs due to precise promotion calibration.

Visual search and size compatibility. CLIP embeddings for image search — deploy in 2–3 weeks: clip‑ViT‑B‑32 or clip‑ViT‑L‑14, Faiss or Qdrant index, REST API. For size recommendation — specific models on return data and reviews with fit indication.

Manufacturing: Quality Inspection and Predictive Maintenance

Quality control and defect detection. CV models for product inspection are one of the most mature industry tasks. YOLOv10 for defect detection, SegFormer for segmentation. Specifics: class imbalance (defects are rare), high recall requirement (missing a defect is worse than false alarm). Typical dataset: 500–2000 defect images + 500–1000 normal. Few‑shot learning via DINO or SAM 2 works with 50–100 annotated examples. We gained experience on an electronics production line — recall 0.95 at FPR 0.03. A predictive maintenance deployment saved a manufacturing client $500,000 per year in unplanned downtime.

Predictive maintenance. Vibration sensors, current sensors, thermocouples → feature extraction → anomaly or mode classification. Models: LSTM‑AE for unsupervised, LightGBM for supervised (if failure history is available). Integration with SCADA/OPC‑UA via opcua-asyncio or MQTT. Key metric: False Negative Rate — a missed pre‑failure is more costly than a false alarm. Threshold tuned to business cost of each error type. Timeline: 3 to 6 months to production.

Digital twin and simulation. Surrogate models — ML models replacing expensive physical simulation. If a CFD simulation takes 6 hours and a surrogate (trained on 10,000 simulations) takes 0.01 seconds, that's 2,000,000× speedup for optimization. SALib for sensitivity analysis, botorch for Bayesian optimization on top of surrogate.

General Principles of Industry AI

Regardless of industry, there are patterns that work everywhere. Data matters more than architecture. In healthcare, 1000 quality labeled images are better than 100,000 poor ones. In manufacturing, 200 real defect examples are more valuable than 10,000 synthetic ones. Compliance‑first design — regulatory requirements are easier to embed into architecture from the start than to add later. Logging, explainability, versioning from day one. Domain expert on the team — an ML engineer without domain knowledge does slowly and error‑prone what an ML engineer plus a doctor/financier/technologist does quickly and correctly.

We guarantee certification to customer requirements (ISO 13485, SOC 2, GDPR) and provide full model documentation (model card, datasheet, compliance report). Our experience: 10,000+ engineering hours and 80+ projects.

Work Process for an Industry AI Solution

1. Domain immersion (2–3 days) — interviews with experts, studying regulatory requirements, auditing available data.
2. MVP design (1–2 weeks) — stack and architecture selection, feasibility assessment.
3. Development and validation (from 4 weeks to 6 months depending on industry) — model training, testing, compliance.
4. Integration and deployment (1–4 weeks) — on‑premise / cloud / edge, documentation, staff training.
5. Support and monitoring — model drift, retraining, SLA.

Estimated timelines:

Cost is calculated individually for each project. Get a consultation — we will evaluate your dataset, regulatory map, and business goals.

Why Choose Our Industry AI Solutions?

• 80+ completed projects in fintech, healthcare, retail, and manufacturing.
• 5 years on the market — proven experience with compliance and deployment.
• Quality guarantee: we ensure target metrics (AUC, recall, latency p99) and provide full documentation.
• Licensed technologies: PyTorch, MONAI, LightGBM, Qdrant — we use open‑source with commercially safe licenses.
• Flexibility: we work as a contractor or as an extension of your team.

Contact us for a free data audit and consultation. Request a proposal with a detailed work plan. We will discuss your task and prepare a commercial proposal.