QuantPi Joins NVIDIA Halos AI Systems Inspection Lab Ecosystem to Advance Trustworthy Physical AI
Read announcement

Reliable Object Detection

QuantPi tests object detection systems for safety-critical applications: pedestrian detection in autonomous vehicles, industrial vision, and visual surveillance. These models predict bounding boxes around each instance of a target class in image or video streams. Because their outputs drive decisions with physical consequences, silent failure modes hidden by aggregate accuracy must be surfaced before deployment.

Typical Failure Modes

What aggregate accuarcy doesn’t tell you.

Environmental Drift

Silent degradation under environmental conditions outside the training distribution (adverse weather, low light, motion blur).
View modalities

&

Critical Edge Cases

View modalities

&

False negatives on safety-critical outliers such as heavily occluded subjects, unusual poses, rare actor types.

Subgroup Disparities

View modalities

&

Performance differences across subject attributes that aggregate metrics conceal.
Research Validation

A glimpse into the QuantPi platform

QuantPi is in continuous exchange with world-leading researchers in machine learning, data science and mathematics. Having emerged from fundamental research, we are convinced that this is a fundamental prerequisite to ensure assessments of AI systems are trustworthy themselves.

Image Classification
Object Detection
Text Generation
LLM Faithfulness
Hallucination Detection
Robustness under Input Perturbations
Image Classification
Object Detection
Text Generation
LLM Faithfulness
Image
Text
Tabular
Time-Series
Video
Image to Text
Image
Text
Tabular
Time-Series
Video
Image to Text
LLM
RAG
Classification Model
Deep Learning Model
LLM
RAG
Classification Model
Deep Learning Model
LLM
RAG
Classification Model
Deep Learning Model
LLM
RAG
Classification Model
testing approach

How QuantPi valdiates object detection systems

Domain InformationDimensional DecompositionAcceptance CriteriaDeployment Decision

Domain information

Every assessment starts from an explicit operational design domain (ODD): target classes, deployment environment, camera configuration, operational scenarios.
View modalities

&

Dimensional Decomposition

The safety- and quality-relevant dimensions derived from the ODD cluster into:
‍
- Subject attributes (age, gender, ethnicity, pose, clothing)
- Subject visibility (occlusion, scale, 3D position)
- Scene context (indoor/outdoor, surrounding objects)
- Environmental conditions (weather, time of day, illumination)
- Image acquisition (brightness, sharpness, resolution)
- Robustness to perturbations (noise, motion blur, synthetic weather and lighting shifts)
View modalities

&

Acceptance Criteria

Acceptance criteria are predefined, measurable performance thresholds the system must meet on each tested dimension to qualify for deployment. In object detection use cases, the most common acceptance criteria is recall at a defined IoU threshold, measured per dimension and sub-group, with stricter thresholds on safety-critical sub-groups.
View modalities

&

Deployment Decision

Models meeting all acceptance criteria are cleared for deployment; partial failures localise where data augmentation, retraining, or post-processing mitigation is required.
View modalities

&

Driving Real-World Impact with Trusted AI

Real-world examples of how companies use QuantPi to build trustworthy AI — from identifying weaknesses to achieving reliable, production-grade performance

Car damage detection for insurance claims

Computer vision model to analyze vehicle images to identify and segment damage, enabling automated insurance coverage estimation to customers.

Client/Partner

ControlExpert

Challenge

Ensuring the robustness of a car damage segmentation model for diverse real-world conditions such as weather, time of day, car pose) is difficult and requires costly manual data annotation.

Assessment Scope

The QuantPi platform was leveraged to rigorously test the model’s performance in depth and continuously allow for improvements to the computer vision model in a matter of hours, instead of weeks. 

Metrics

Automated weather perturbation testing (e.g. day/night, clear/rain) was implemented using Stable Diffusion and CLIP models to identify outdoor images.

Data annotation was automated by using a CLIP model to dynamically tag images with relevant attributes (e.g. car pose, indoor/outdoor, raindrops).
‍
Model performance was evaluated by using IoU across tagged attributes to identify weaknesses and guide targeted improvements

Business Impact

- Reduced manual effort for Data Scientist
- Increase model performance
- Reduced claims processing time

RAG & LLM for Factory Troubleshooting in Life Sciences

LLM and RAG based AI system for factory troubleshooting. Assisting line workers in identifying the root cause to issues quickly and enabling speedy resolution.

Client/Partner

Violet AI
Read the entire success story

Challenge

Violet AI needed to demonstrate the quality and reliability of their RAG-based AI systems, particularly regarding document retrieval accuracy and response faithfulness, to meet compliance demands in the life sciences industry, including the EU AI Act.

Assessment Scope

The QuantPi platform was leveraged to address these challenges, rigorously evaluating the RAG and LLM systems' performance and preparing Violet AI for compliance.

Metrics

We evaluated RAG system quality, incorporating metrics for retrieval accuracy (LAcc, SAcc, MRR) and LLM faithfulness/hallucination (based on RAGAs framework and existing out of the box metrics from QuantPi).

Business Impact

- Reduced manual effort for Data Scientist
- Increase model performance
- Reduced claims processing time

AI-Powered Investment Research Assistant

Agentic workflow using LLMs as an orchestrator for summarization, multi-turn chatbots, and data extraction from unstructured sources.

Client/Partner

Unique
Read the full case study

Challenge

The system faced significant risks regarding the failure to meet regulatory requirements or internal company guidelines (e.g., recommending restricted products). Other critical concerns included the potential oversharing of client data, inconsistent advice quality across different customer segments, and the risk of poor financial outcomes

Assessment Scope

QuantPi provided a comprehensive, black-box AI testing technology to evaluate the model for performance, bias, robustness, and data quality. The pilot testing focused specifically on three core areas: Accuracy, hallucination and reliability.

Metrics

Hallucination Risks: Measured via a faithfulness metric to ensure responses were grounded in the provided context.
Document Search Risks: Assessed using Word Overlap Rate, Mean Reciprocal Rank (MRR), and Lenient Retrieval Accuracy.
Reliability: Tested across various query difficulty levels, domain bias, and typo tolerance.

Business Impact

- Risk Mitigation: Identified and addressed risks related to advisor misuse and inadequate clarity in investment recommendations.
‍
- Performance Assurance: Guaranteed the grounding of AI-generated advice in provided financial contexts to prevent misinformation.

CV Matching System

StepStone develops an LLM-based candidate recommender system which automatically provides recruiters with candidate recommendations purely based on job listings.

Client/Partner

StepStone, TÜV AI.LAB
Read the full white paper

Challenge

Stepstone needed to validate and justify the fairness of their LLM-based candidate recommender system, particularly regarding non-discrimination of applicants based on age, gender and ethnic origin.

Assessment Scope

The QuantPi Platform was used to assess false negative and selection rates of the system with regards to the sensitive attributes and their perturbations – resulting in quantitative test results on non-discrimination on individual and group level.

Metrics

An assurance case was built together with TIC partner TÜV AI.Lab. This document serves as structured line of argumentation and links the quantitative evidences generated in the QuantPi Platform to applicable legal requirements under Germany’s Equal Treatment Act (AGG) and future provisions of the EU AI Act.

Business Impact

- Reduced bias and discrimination
- Increased system performance
- Compliance documented in assurance case

QuantPi's evaluation of object detection systems produces:

Granular Performance Analysis

A per-dimension breakdown across every defined sub-group and metric, surfacing silent failure modes.

Failure Diagnostics

A diagnostic identifying where the model underperforms its acceptance criteria.

Audit-Ready Evidence

A traceable, repeatable evidence package supporting deployment decisions and external audit requirements.

Data Quality Insights

A data quality breakdown, surfacing uncertainty with respect to measurements. Potentially informing where more test data is needed to be more confident.

See QuantPi's continuous robustness assurance for a computer vision model in automotive claims processing

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Request Demo
Request Demo
Request a Demo