Organization · Venture Capital

Irregular Expressions

Irregular Expressions is an early-stage venture capital fund based in San Francisco and New York, focusing on pre-seed and seed investments in startups developing infrastructure software, AI, and developer tools. The firm leverages a network of over 90 engineering leaders to support its portfolio companies.

Venture CapitalFunder

Overview

What Irregular Expressions does

What Irregular Expressions is

Irregular Expressions is an early-stage venture capital fund based in San Francisco and New York, focusing on pre-seed and seed investments in startups developing infrastructure software, AI, and developer tools. The firm leverages a network of over 90 engineering leaders to support its portfolio companies.

How it participates in the startup ecosystem

The current public profile supports funder capabilities. Each capability below is shown separately so organization identity and role-specific information do not get conflated.

Operational profile

Organization status and context

Organization status describes the entity itself. Funding and program availability appear within the relevant capability profile.

Operational status
Active
Organization type
Venture Capital
Structured capabilities
Funder
Last reviewed
May 1, 2026

Official channels

Find Irregular Expressions elsewhere

Open the official website or a supported primary organization account.

Capability profile

Irregular Expressions as a funder

Investment thesis

Irregular Expressions invests in and advises early-stage startups focused on infrastructure, data engineering, and AI. The firm supports technical founders building products for engineering, platform, and security teams, leveraging a network of senior engineering leaders to provide guidance.

Focus narrative

Irregular Expressions primarily invests in pre-seed and seed-stage companies that are developing infrastructure software. The firm emphasizes sectors including DevOps, AI, machine learning, data engineering, security, observability, and developer productivity, with a mission to empower startups that enhance the capabilities of engineering teams.

Value beyond capital

Irregular Expressions offers significant value to its portfolio companies through access to a network of over 90 senior engineering leaders and operators. This network acts as an ongoing resource for product development, engineering insights, and go-to-market strategies, helping startups validate their roadmaps and enhance enterprise adoption.

Focus and fit

Stages
Pre-seed · Seed
Sectors
AI · Devops · Observability

Availability

Currently deploying
Yes

Portfolio and track record

Recorded notable exits
Canva and People.ai

Portfolio context

Notable portfolio companies include: 4D Sight: Creates virtual billboards for e-sports leagues, enhancing audience engagement without disrupting gameplay. Bubble: A no-code software development platform that allows users to design and launch applications without programming skills. People.ai: Provides a platform that transforms business activity data into actionable insights for sales, marketing, and operations. Canva: An online design and publishing platform enabling users to create professional-quality designs.

Common questions

What to know about Irregular Expressions

Upseed classifies Irregular Expressions under the Venture Capital organization type. Its current public profile includes funder capabilities.

Next step

Keep comparing startup organizations

Return to the organization index or continue through a verified sector, place, peer, or program connected to this profile.

Source and update disclosure

Know what this page is based on

Source basis

Public organization and program material, official websites, and structured Upseed records.

Last reviewed

May 1, 2026

Known limits

Capability profiles appear only when a structured public profile also contains decision-useful role evidence. Terms and availability can change.

Upseed is an independent guide. It is not affiliated with or endorsed by the organization described on this page.