The Virtual Scientist
for Spectrometry

Spectral molecular profiling is too complex to scale.
We make it autonomous.

Spectral molecular profiling is too complex to scale.

Spectrometry is the foundational technology for modern analytical science. But today’s workflows remain highly manual, expert-dependent, and difficult to scale.

Mass spectrometer

Powerful instruments

Modern spectrometers combine extraordinary sensitivity and resolution, allowing scientists to measure thousands of molecules across complex biological and environmental samples.

Instrument parameter controls

Fragile configurations

Powerful instruments still depend on fragile configurations. Choosing the right chromatography gradients, collision energies, and monitored transitions can take weeks, and each method must be re-optimized across instrument models, sample types, and experimental conditions.

Chromatogram signal

Data processing bottleneck

From peak detection and integration to quantification, turning raw spectrometry signals into reliable quantitative data remains highly manual, operator-dependent, and difficult to reproduce at scale.

Scientist operating analysis software

Untapped downstream potential

The complexity of spectrometry workflows still limits the reliability and scalability of molecular profiling across healthcare, biotech, pharma, CROs, and environmental testing.

Precision Medicine From metabolomics and proteomics to oncology biomarkers, pharmacokinetics, and clinical toxicology, spectral molecular profiling is becoming essential to modern medicine.

Consumer & Environmental Chemistry New regulatory pressure is creating urgent demand for scalable analytical testing in cosmetics, food, environmental, and consumer product sectors.

We make spectrometry autonomous.

The Virtual Scientist for Spectrometry

MendAI Labs is building an AI agent, connected to scientific knowledge, that interacts with spectrometers in real time: configuring instruments, analyzing results, deciding when additional runs are needed, and quantifying the molecules present in complex samples.

VIRTUAL SCIENTIST PROPRIETARY KNOWLEDGE BASE
H
HELIX
refines configuration
G
GENESIS
1st initialization
Instrument
runs analysis
P
PRISM
scores results
PRISM PRISM

Peak Integration that Learns

Turns raw spectrometry signals into reliable molecular concentrations.

  • Analyzes entire batches, not isolated samples, to correct batch effects and instrument drift
  • Learns locally from expert feedback while keeping data on-site
  • Automates up to 90% of signal processing
HELIX HELIX

Real-Time Control Engine

Continuously optimizes analytical methods through autonomous instrument control.

  • Uses Bayesian optimization to dynamically adjust acquisition parameters in real time
  • Refines experimental methods through closed-loop feedback from instrument results
  • Turns method development into a self-optimizing process
GENESIS GENESIS

AI-Driven Method Design

Turns biological questions into sample preparation protocols and analytical methods.

  • LLM-based agents connected to the scientific literature and manufacturers' application notes
  • Accelerates initial method design from days to minutes
  • Starts from the biological question, not manual expert configuration

Prism | Peak Integration that Learns

MendAI's machine learning-powered peak integration software
for LC-MS/MS and GC-MS workflows.

Unlike traditional rule-based software, Prism learns from operator corrections and continuously improves the reliability of peak integration across analytical batches.

→  Fewer manual corrections
→  Faster analytical workflows
→  Higher reproducibility
→  Scalable spectrometry operations

An Interdisciplinary Founding Team

MendAI Labs began when Guillaume and Paul met at Mila in 2025, united by a shared conviction: meaningful impact on patient health requires multidisciplinary expertise and a nuanced understanding of clinical needs. View our publications →

Paul Bertin

Paul Bertin

PhD | Co-founder & CEO

Mila · Université de Montréal · École polytechnique Paris

Completed his PhD at Mila under the supervision of Prof. Yoshua Bengio, focusing on modelling cellular responses to biological perturbations. His work included an iterative experimental design approach prospectively validated with a chemical biology team, as well as an interpretable generative model for large-scale omics data.

The scientific foundations of MendAI’s AI solutions build directly on this expertise at the intersection of deep learning and biology.

Guillaume Bachelot

Guillaume Bachelot

PharmD, PhD | Co-founder & CSO

AP-HP · Sorbonne Université · Mila

Spent eight years as a hospital practitioner at the Paris university hospitals and Sorbonne University. He led translational research initiatives within the spectrometry platform at Saint-Antoine Hospital, and developed innovative diagnostic tests integrating machine learning and spectrometry.

This dual clinical and scientific expertise underpins the medical relevance and real-world applicability of MendAI's solutions.

Scientific & Strategic Partners

Jacques Corbeil

Jacques Corbeil

Professor of Medicine, Research Chair in Medical Genomics

Alexandre Le Bouthillier

Alexandre Le Bouthillier

Member of the Mila Board of Directors, Investor & Entrepreneur

Sorbonne Université

Sorbonne Université

Partner institution for Prism validation & clinical research expertise

Mila

Mila

Incubated by Mila Ventures, with access to offices, compute resources, and expert advisors.

Ready to automate your spectrometry workflows?

Whether you operate a spectrometry platform, work in healthcare, pharma, research, or industrial testing, or are exploring strategic partnerships, we’d love to hear from you.

contact@mendailabs.com