Deep Learning Startups

Akido Labs is a healthcare technology company building AI-based systems intended to make healthcare more anticipatory, precise, and accessible. Founded in 2015, it focuses on applying AI to the delivery of healthcare.

14 investors

Overview.ai builds custom hardware, edge AI, and software systems that give manufacturers real-time visibility into their production processes to catch defects earlier and reduce waste. It serves industrial manufacturers seeking to improve quality control at scale.

12 investors

Activeloop builds a continual learning layer for coding agents that captures agent interactions, codifies them into reusable skills, optimizes those skills, and propagates them across a team's AI coding agents. It works with tools like Claude Code, Codex, and Cursor and stores customer traces and data on the customer's own cloud storage, serving engineering teams building with AI agents.

10 investors

D-ID builds AI and deep-learning technology for generating realistic AI personas, animating still photos, and producing videos through a platform and APIs aimed at creators and developers. It originally created a facial image de-identification solution to protect images and videos from facial recognition software.

10 investors

Mashgin develops automated checkout kiosks using computer vision and machine learning to instantly identify items without scanning or barcodes, serving retailers in stadiums, hospitality, hospitals, and foodservice venues.

9 investors

Ligo develops computational tools for enzyme design by combining large language models with specialized neural networks to enable scientists and AI agents to discover new enzymes and biological functions.

8 investors

Netomi AI builds artificial intelligence for automating customer care and support for large consumer brands. Its technology handles customer engagement and service interactions using generative AI.

8 investors

Anthrogen develops Odyssey, a family of protein language models with 102 billion parameters designed for multi-objective protein design. The models provide an API for researchers and organizations in biotechnology and drug discovery.

7 investors

Numerion Labs develops machine learning algorithms for drug discovery, combining computational chemistry, structural biology, and medicinal chemistry. It builds AI-driven platforms aimed at accelerating the identification of new medicines.

7 investors

Anima is a healthcare platform that combines online consultation with productivity tools, featuring EHR integration, advanced analytics, and AI-powered document processing to streamline primary care workflows and improve patient outcomes.

6 investors

Cerrion develops AI solutions for the manufacturing industry to improve efficiency, safety, and resilience in production operations.

6 investors

Biodock provides an AI image analysis platform designed for researchers and scientists to train, run, and interpret deep learning models on biological images without requiring coding expertise. The platform enables rapid model creation, distributed execution at scale, and detailed quantitative analysis for research organizations in biotech and life sciences.

5 investors

NDEA develops artificial general intelligence technology and benchmarks. The company is founded by the creator of Keras and a co-founder of Zapier.

5 investors

Automorphic develops machine learning infrastructure where deployed models continuously improve themselves using generated data as training material.

4 investors

Cedana provides live migration technology for CPU and GPU containers that saves workload state, migrates jobs across nodes or clusters, and resumes execution without code changes. The platform serves compute infrastructure teams and enables workload portability across Kubernetes, SLURM, AWS, Google Cloud, and NVIDIA environments.

4 investors

Diffuse designs proteins using AI models trained on experimental data collected through proprietary assays and software, delivering validated AI-designed proteins to partners.

4 investors

Expanse predicts the GPU, memory, runtime, and failure risk of compute jobs before they run, so clusters do more work on the same hardware and reduce failed or overprovisioned jobs. It works alongside existing schedulers such as SLURM, Kubernetes, or Nomad, serving engineers running training, HPC simulations, and inference workloads.

4 investors

Focal Systems builds AI and deep-learning retail automation tools that digitize store shelves and run an operating system for brick-and-mortar retailers to optimize merchandising, operations, and supply chains. It serves physical retail store operators and management teams.

4 investors

Jido Maps helps teams, with or without machine learning expertise, turn their visual or scanned data into deployed computer vision models for applications such as security camera monitoring, data entry automation, image interpretation, and augmented reality. It offers to deliver an initial proof of concept within a week.

4 investors

Mindee provides an AI-powered OCR API that extracts structured data from documents including invoices, receipts, passports, and other formats, converting unstructured documents into JSON with high accuracy. The platform serves businesses across industries that need to automate document processing workflows without requiring model training.

4 investors

Vathys worked on accelerating machine learning and deep learning through software, after initially building machine learning hardware intended to be faster than ASICs. The company operated in the AI/ML infrastructure space.

4 investors

Flywheel AI provides remote operation systems for excavators and construction equipment, enabling operators to control multiple machines or sites from a distance while improving site safety and productivity.

3 investors

Induction Labs is a research lab developing foundation models designed to learn from their own observations and update over time, with the goal of scaling curiosity-driven artificial intelligence toward superintelligence. It positions itself in the AI research and foundation model space.

3 investors

Rasyn.ai offers Marigold, an AI research assistant that lets scientists ask research questions, run tools, or delegate entire tasks, returning results such as structures, charts, code, and papers in panels. It provides 54 runnable tools, 40 executing on connected hardware and 14 on hosted endpoints.

3 investors