Activeloop
Series A in 2024
Activeloop is a technology company that specializes in providing a distributed cloud computing platform designed for building and scaling data pipelines for machine learning applications. Its flagship product, Deep Lake, is a database optimized for artificial intelligence, particularly for deep learning and large language model (LLM) applications. Deep Lake offers a unique storage format that accommodates various data types, including embeddings, audio, text, videos, images, and annotations. The platform features capabilities for querying, vector search, data streaming during model training, and comprehensive data versioning and lineage management. Deep Lake is designed to handle data of any size in a serverless environment, facilitating seamless integration with popular tools such as LangChain and Weights & Biases. The platform is utilized by notable organizations, including Bayer Radiology and Matterport. Founded by a team with backgrounds from prestigious institutions such as Princeton and Stanford, Activeloop has garnered support from Y Combinator, positioning itself as a key player in the machine learning and AI data management landscape.
AI21 Labs
Series C in 2023
AI21 Labs is an Israeli artificial intelligence company that develops generative AI and large language models along with tools to understand and generate natural language. Founded in 2017 and based in Tel Aviv-Yafo, it provides AI-based writing and reading assistance that understands context and meaning, including a writing companion that can rephrase text to convey the intended message and generate summaries. The company's tools help users read more efficiently, analyze and visualize complex knowledge structures, and apply deep learning and NLP to transform writing and comprehension for businesses and individuals.
MosaicML
Venture Round in 2023
MosaicML is a company focused on creating an efficient infrastructure for training large language models and improving the overall efficiency of neural networks. It develops software and artificial intelligence training algorithms that enhance the training process by utilizing algorithmic techniques such as sparsity and network pruning. These innovations allow users to effectively and securely train large-scale AI models on their proprietary data, while also optimizing for speed, quality, and cost. MosaicML aims to streamline the machine learning model recomposition process, making it easier for organizations to harness the power of AI in their operations.
HelperAI
Pre Seed Round in 2022
HelperAI lets enterprises build a 24/7 AI workforce, automating routine tasks, and breaking language barriers for global expansion. With seamless integration with your business SOPs, CRM systems, and popular tools, our AI agents redefine customer support, sales, and marketing functions. Experience unprecedented efficiency with HelperAI.
Founded in 2017, Verbit specializes in AI-driven transcription and captioning services. It offers solutions for various industries, including education, legal, media, and enterprise, making audio and video content searchable, accessible, and actionable.
Unbabel Inc., established in 2013, specializes in online translation services. Headquartered in San Francisco with additional offices in New York, Pittsburgh, and Lisbon, the company combines machine translation with human editing to facilitate global business communication. Unbabel translates user-generated content like emails, wikis, and forums, offering solutions for customer service, tickets, and chat. Its services enable enterprises to serve customers in their native languages, bridging language barriers and fostering multilingual understanding.
Unbabel Inc., established in 2013, specializes in online translation services. Headquartered in San Francisco with additional offices in New York, Pittsburgh, and Lisbon, the company combines machine translation with human editing to facilitate global business communication. Unbabel translates user-generated content like emails, wikis, and forums, offering solutions for customer service, tickets, and chat. Its services enable enterprises to serve customers in their native languages, bridging language barriers and fostering multilingual understanding.
Kngine
Venture Round in 2014
Kngine, Inc. is an intelligent assistant and answer engine that provides users with direct answers to their inquiries. Founded in 2008 and headquartered in San Mateo, California, with an engineering office in Cairo, Egypt, Kngine is designed for developers and businesses to create smart applications. Its platform incorporates advanced natural language understanding and utilizes deep learning, big data, and knowledge-based approaches to deliver accurate responses. Kngine’s technology extracts factual information from web content to enhance its knowledge graph, allowing it to comprehend and process user questions effectively. The platform is capable of automating customer service, powering voice interfaces, and enhancing enterprise search functionality, ensuring that users receive meaningful results. Kngine also offers a beta application for iOS and Android devices, further extending its accessibility. As of 2017, Kngine operates as a subsidiary of Samsung Research America, Inc.
Traceloop specializes in monitoring the performance of large language models (LLMs) in production environments. It offers automated benchmarks, drift detection, real-time alerts, and gradual rollout capabilities to help AI developers and engineering teams ensure reliable deployment of changes to prompts and model configurations.