RamAIn

Super fast computer-use agents to build AI employees.

Winter 2026ActiveB2BArtificial IntelligenceMachine LearningComputer VisionSan Francisco, CA, USA

Company

RamAIn can intelligently automate any super complex and repetitive process across applications (Browser & Desktop) super fast. It doesn’t matter whether you have API access or not - if you can see it, RamAIn can do it. Current CUAs all follow the same decision loop: screenshot, VLM, decision, repeat. This is expensive, unreliable and slow. We build infrastructure to pre-train CUAs on specific interfaces, effectively teaching RamAIn to use your computer.

Verdict

High Signal

Market Opportunity

Computer-use / RPA automation is a multi-billion dollar market (UiPath alone is $10B+ company). Clear B2B ICP: mid-market/enterprise teams stuck on legacy systems, EHRs, ERPs, vendor portals with no API access. Monetization path is obvious — per-seat or usage-based 'AI employee' pricing. Insurance, healthcare, and finance verticals are high-value with demonstrated willingness to pay for automation.

High Signal

Founder Signal

Shourya (CEO) has prior entrepreneurial traction: built Genoshi (enterprise AI studio) to six-figure revenue, plus sold a 50k-user gaming community. Ex-McKinsey on enterprise AI in insurance — directly relevant to target ICP. Vansh (CTO) is an unusually strong technical founder: published at ICLR 2025 and ACS, built PANORAMA (merged into Meta's FAISS), researched at CMU MLD and University of Copenhagen. Both are IIT Delhi CS/EE; Vansh dropped out to join YC. Relatively young (22) but concrete prior output is well above average for their age.

Low Signal

Competition

Extremely crowded: Anthropic Computer Use, OpenAI Operator, MultiOn, Browserbase, Playwright-based agents, UiPath, Automation Anywhere all compete in adjacent or direct space. RamAIn's differentiation claim — pre-training on specific UIs to build a 'Visual DOM' instead of screenshot→VLM loops — is technically interesting but unproven at scale. Big tech (Anthropic, OpenAI, Google) are actively investing here and could commoditize the core approach quickly.

Medium Signal

Product

Live website with playground demos, feature descriptions, and named integrations (Salesforce, SAP, Oracle, QuickBooks). Founder posts confirm real customers and 'automation running in production 24x7' with three SMB/MSE design partner use cases (insurance underwriting, data pipelines, pricing extraction). No public revenue numbers, named customer logos, or pricing page — still early but beyond pure vaporware.

OverallB Tier

RamAIn has two genuinely strong founders — a CEO with proven revenue from a prior AI studio and a CTO with Meta FAISS contributions and ICLR publications, both at IIT Delhi. The early design partner traction in insurance underwriting and vendor portals is real and the ICP (legacy/API-less enterprise systems) is well-defined with high willingness to pay. The fatal risk is competitive: this is one of the most crowded categories in AI right now, with Anthropic, OpenAI, and a dozen well-funded startups attacking the same problem. The 'pre-train on specific UIs' angle is a reasonable technical wedge but needs to demonstrate durable differentiation before the big players catch up. Worth watching, but needs to show defensibility fast.

Active Founders

Shourya Vir Jain

Co-Founder & CEO

Entrepreneur building computer use agents to power the next generation of workflow automation. EE @ IIT Delhi, ex-McKinsey & Company where he worked on enterprise AI projects in insurance. Built and ran Genoshi, an enterprise AI studio delivering RAG pipelines and workflow automation systems that scaled to six-figure revenue. FIDE-rated (2118) competitive chess player who has represented India across 17 countries, and previously built and sold a 50k-user gaming community.

Vansh Ramani

Co-Founder & CTO

Entrepreneurial researcher building fastest computer use agents. CS @ IIT Delhi. Published at ICLR and ACS -- work cited 20+ times including in pharmaceutical research, NeurIPS, and Nature journals. Built PANORAMA, the fastest nearest-neighbor search algorithm, merged into Meta's FAISS. Former researcher at CMU's MLD and University of Copenhagen. Research funded by the Danish Data Science Academy and Pioneer Centre for AI. National robotics champion at Lam Research's Engineering Challenge.

RamAIn

TierB Tier

BatchWinter 2026

Team Size2

StatusActive

LocationSan Francisco, CA, USA