Clarifai

At Together AI, we believe software quality comes from research-driven innovation and genuine collaboration. Our engineers and researchers work side-by-side, contributing to open-source advancements like FlashAttention, Mamba, and RedPajama while building infrastructure that powers the next generation of AI. We optimize for impact over hierarchy - meaning you'll own substantial technical challenges from day one, whether it's designing distributed inference engines or pioneering new model architectures. We're committed to open and transparent AI systems because we believe the best outcomes for society emerge from shared knowledge. Our purpose-built GPU cloud platform empowers developers and organizations of all sizes to train, fine-tune, and deploy generative AI models without lock-in. We foster a culture of deep technical curiosity where tackling frontier problems in distributed systems, model optimization, and AI infrastructure is the norm - not the exception.

In 2023, a group of visionary scientists and investors recognized that drug discovery remained an expensive, time-consuming process with high failure rates. Despite advances in AI, the pharmaceutical industry was still struggling to translate computational insights into viable medicines. This realization sparked the creation of Xaira Therapeutics - co-founded and incubated by ARCH Venture Partners and Foresite Labs - to build an integrated biotechnology company that would truly harness the transformative power of artificial intelligence for drug discovery. Built on the groundbreaking protein design work of Nobel Laureate Dr. David Baker and his team at the University of Washington's Institute for Protein Design, Xaira brings together advanced AI research, expansive data generation, and robust therapeutic product development. The company emerged from stealth mode in 2024 with over $1 billion in funding, ready to revolutionize how we treat disease by making predictions about the right biology to target, the best therapeutics to modulate those targets, and the patients who will benefit most. With offices in South San Francisco, Seattle, and London, Xaira is shifting drug discovery from an empirical science to a precise engineering discipline.

LangChain builds the frameworks and platforms that developers use to construct, test, and deploy AI agents and large language model (LLM) systems. Its open source offerings - the LangChain and LangGraph frameworks - together exceed 90 million downloads per month and connect to more than 1,000 integrations, making them among the most widely adopted tools in the agent engineering ecosystem. The company operates on an open-core model, pairing its open source frameworks with LangSmith, a commercial platform that adds observability, evaluation, and deployment capabilities for production LLM systems. The open source frameworks provide pre-built architectures and developer tooling for assembling AI agents, while LangSmith addresses the reliability and monitoring demands that arise when those systems move into production. Together, the products cover the full development lifecycle - from early prototyping through to live deployment - targeting both the individual developer and the AI team operating at scale. LangChain's user base spans major enterprises and early-stage startups alike, with millions of developers relying on its frameworks worldwide. The company is focused specifically on agent engineering: the discipline of building LLM-powered systems that can reason, plan, and act reliably in real-world environments.

Duku AI builds an autonomous testing platform designed to remove the manual burden of software quality assurance. Its platform deploys AI agents that simulate real customer journeys after every build, catching failures before users encounter them. Rather than requiring teams to write and maintain test scripts, the system self-heals as codebases evolve - eliminating the flaky tests and stale test suites that slow down engineering teams. The company is led by operators with experience scaling Meta's testing infrastructure, launching Uber's global playbooks, and growing Deliveroo from its early stages to hypergrowth. That background informs both the product's technical architecture and its positioning: giving engineering and product teams the confidence to ship quickly without sacrificing reliability. Despite being a three-person team, Duku AI has reached $330,000 in revenue in 2025. The company is venture-backed.

Descript makes a video and audio editing platform built around a straightforward idea: editing should work like editing a document. Rather than working with traditional timelines, users edit by manipulating a transcript - cutting text cuts the corresponding audio or video. The platform supports solo creators and distributed teams, and claims a user base of millions. The product spans video editing, audio editing, podcast production, and collaboration tools, with an expanding suite of AI-assisted features. These include automated design and the ability to generate content from a text description. The technical domains the company works across include generative AI, speech processing, and human-centered design - with a stated emphasis on building AI that augments rather than complicates the creative process. Descript describes its team as drawing from both the creator world and the technology industry. The company places weight on human-centered design and positions itself against the complexity of conventional editing software. Its culture signals point toward a team that values invention and forward-looking thinking - people, as the company puts it, who think about things that don't yet exist.

In 2019, Edo Liberty was working as a research director at AWS and at Yahoo before that, where he witnessed the tremendous power of combining AI models and vector search to dramatically improve applications such as spam detectors and recommendation systems. While building custom vector search systems at enormous scales, he assumed there was already a packaged solution available for everyone else who didn't have access to the same engineering and data-science resources. To his surprise, there wasn't. This gap in the market led Edo to found Pinecone, creating the vector database category of solutions from scratch. The company was born from the recognition that developers and engineering teams of all sizes needed accessible storage and retrieval infrastructure for building and running state-of-the-art AI applications. This founding principle of accessibility drove Pinecone's evolution into a fully managed service known for its ease of use, enabling thousands of companies to ship AI applications faster and more confidently.