Skip to main content

CentML Inc.

Category: AI Infrastructure

CentML provides advanced software tools to optimize and accelerate large-scale machine learning model training and inference for enterprises. CentML Inc. was founded in 2022. The company is led by Gennady Pekhimenko. Based in Toronto, Ontario, Canada. Team size: 30-50. Total funding raised: $30.9 million. Latest round: Seed Round. Key investors include Gradient Ventures (Google), Radical Ventures, NVIDIA NVentures, Deloitte Ventures.

Founded
2022
Headquarters
Toronto, Ontario, Canada
Team size
30-50
Total funding
$30.9 million

Value proposition

Dramatically reduces cost and time required for training/inference by optimizing how models run on existing hardware—without requiring changes to model code or architecture.

Products and solutions

CentML Platform/Core Optimizer, Inference Acceleration Tools, Cost Reduction Analytics Dashboard

Unique value

CentML’s core innovation is its ability to transparently optimize deep learning workloads across different hardware backends—improving performance/cost without requiring manual tuning or code changes by data scientists/engineers.

Target customer

Enterprises building or deploying large-scale AI models—especially those seeking to reduce compute costs or improve efficiency on cloud/GPU infrastructure.

Industries served

Generative AI / Large Language Models (LLMs), Computer Vision, Finance, Healthcare AI research labs

Technology advantage

Compiler-level optimizations tailored for modern GPUs/cloud accelerators. Automated workload analysis that identifies inefficiencies invisible to standard profiling tools. Seamless integration with existing ML pipelines/frameworks (“drop-in” solution).

How they differentiate

Unlike some competitors who focus solely on deployment/inference or require significant workflow changes, CentML emphasizes transparent integration into existing workflows plus both training and inference acceleration.

Main competitors

MosaicML, OctoML, Modular, Weights & Biases

Key partnerships

University of Toronto Machine Learning Group, Leading cloud providers, Nvidia (Acquisition)

Notable customers

Names under NDA but includes “leading global technology firms”, Several Fortune Global top-ten customers cited anonymously

Major milestones

Raised $27M Seed Round. Recognized as one of Canada’s fastest-growing deeptech startups Acquired by Nvidia.

Growth metrics

Not publicly disclosed—but reportedly achieved rapid customer adoption among top-tier tech companies running large-scale LLM workloads.

Market positioning

Positioned as an “invisible accelerator” enabling enterprise teams to maximize ROI from expensive GPU/cloud investments without sacrificing flexibility or accuracy.

Geographic focus

Global

Patents and IP

Specific patent filings are not public yet but their IP centers around proprietary optimization algorithms/software stack.

About Gennady Pekhimenko

Gennady Pekhimenko is the Co-Founder and CEO of CentML, a company that optimizes machine learning models for enterprises. He is also an Associate Professor at the University of Toronto and a Canada CIFAR AI Chair at the Vector Institute. His research focuses on computer systems, machine learning, and hardware acceleration.

Official website: