vision-language-models
(10)
Varun Mathur
Varun Mathur is the co-founder and CEO of Unisson, a Y Combinator Winter 2026 startup building AI agents that act as hands-on product experts for customer-facing B2B teams. Before Unisson he led product, engineering, growth and vision-language-model research for agent products at Ambient.ai (YC W17), where he built a video-understanding model that outperformed Gemini 2.5 Pro and sold AI to Fortune 5 companies. He studied computer science and statistics at the University of Illinois Urbana-Champaign and works out of San Francisco.
OnDeck AI
OnDeck AI is a Vancouver-based Y Combinator (S25) company building Perception-0, a vision model that lets organizations find any object, behavior, or event in video footage using plain language - without collecting training data or building a custom computer-vision pipeline. Started as a fisheries-monitoring project, it now serves defense agencies, robotics teams, and security operators, including a pilot with Singapore's Ministry of Defence.
Unsiloed AI
Unsiloed AI builds an API that turns messy, multimodal documents - PDFs, scans, spreadsheets, charts and images - into structured JSON and LLM-ready Markdown. Founded by IIT Kharagpur alumni Aman Mishra and Adnan Abbas and backed by Y Combinator (Fall 2025), the company uses proprietary dual-stream vision models to parse, extract and split documents at scale, and ranks #1 on the public olmOCR-Bench, ahead of models from OpenAI, Anthropic and Google.
H3 Zoom
H3 Zoom is a Singapore-based deep-tech company that turns drone, robotics, 360-degree and existing visual data into regulator-defensible building and infrastructure inspection reports. Its AI platform automates facade and interior condition assessments - detecting, grading and mapping defects - to cut inspection time and cost while improving worker safety. Spun out of aerospace group H3 Dynamics and founded in 2016, the company has completed more than 10,000 inspections across seven countries and closed an oversubscribed USD $3.6 million Series A in 2026 to scale across Asia.
Liquid AI
Liquid AI is an MIT CSAIL spinout building Liquid Foundation Models (LFMs), a family of general-purpose AI models engineered to run fast and privately on-device - phones, laptops, wearables, and cars - instead of relying on the cloud. Rooted in the founders' 'liquid neural network' research inspired by the nervous system of a roundworm, the company pairs a first-principles model architecture with LEAP, a developer platform that deploys models into mobile apps in a few lines of code, and Apollo, a fully local AI chat app. It raised a $250M Series A led by AMD in December 2024 at a roughly $2.35B valuation.
Lumana
Lumana is an enterprise AI video security company that turns ordinary IP cameras into real-time monitoring, detection and response systems. Founded in 2021 by former Intel computer-vision leaders and based in Los Gatos, California, Lumana pairs its proprietary VIA-1 vision model with vision language models and agentic AI to run analytics on more than a billion images a day - flagging weapons, fires, falls and unauthorized access, cutting false alerts, and letting security teams search footage across 100+ attributes. The company raised a $40M Series A in 2025, bringing total funding to $64M, and counts McDonald's, Meta, the Minnesota Twins and NYU among its customers.
Reducto
Reducto is a San Francisco AI company that turns messy, complex documents into clean, LLM-ready data. Its agentic document platform pairs traditional computer vision with vision-language models and a multi-pass 'Agentic OCR' correction loop to parse, split, extract, classify, and even edit PDFs and other formats through a developer API. Founded in 2023 by MIT grads Adit Abraham and Raunak Chowdhuri, Reducto has processed billions of pages for AI teams and enterprises in finance, healthcare, legal, insurance, and compliance, and raised $108M total including a $75M Series B led by Andreessen Horowitz in October 2025.
Exactrx
Exactrx is a Nashville- and New York-based healthcare AI company that automates revenue cycle management for ambulatory surgery centers and outpatient providers. Its platform reads the operative note, codes the case, validates each charge against more than 80,000 payer rules with a tri-engine system, and generates a submission-ready claim in under 30 seconds - leaving humans to review and submit. Founded in 2024 by Athena Doshi and a clinician-and-engineer team, the company positions itself as access-to-care infrastructure rather than billing software, aiming to recover the revenue that denials and documentation gaps quietly drain from surgical facilities.
Mbodi AI
Mbodi AI is a New York robotics-software company building an embodied AI platform that lets people teach industrial robots new skills through plain natural language and short demonstrations - no code and no specialized engineers. Founded in 2024 by two former Google Public DNS engineers, the platform combines generative AI, agent orchestration and symbolic reasoning to turn everyday instructions into precise, reliable robot actions that run in production within minutes and transfer across an entire robot fleet. Mbodi is a Y Combinator company with a joint commercialization partnership with ABB Robotics.
Sebastian Peralta
Sebastian Peralta is a physicist-turned-roboticist and co-founder of Mbodi AI, a New York startup teaching industrial robots new skills through plain language and quick demonstrations instead of weeks of reprogramming. A former core engineer on Google Public DNS (8.8.8.8) with a triple major in electrical engineering, computer science, and physics from UPenn, he builds embodied-AI systems where one robot's lesson instantly becomes the whole fleet's. Mbodi won ABB Robotics' global AI Startup Challenge, reached the Startup Battlefield Top 20 at TechCrunch Disrupt 2025, and is rolling out factory deployments with a Fortune 100 consumer-goods partner.