Trust and evaluation
QA, observability, and controls for AI-produced work.
Formation · operating support · early capital
We form and back applied AI companies.
Founders, operators, advisors, customers, and angel capital working together from formation to seed.
Investment thesis
We back the layers that make applied AI dependable.
Trust, workflow ownership, and outcome-based services.
Vertical workflows
Software that owns a costly, recurring business process.
AI-native services
Outcome-led firms where AI changes delivery and margin.
Evidence before scale.
Flowleads and Devtx expose real demand. We form companies where the same problem repeats.
Founder program
From product to first market.
A 12-week company sprint for product, go-to-market, customers, and the next financing milestone.
Portfolio
Our companies.
A focused portfolio built from problems we know firsthand.
01 Building
Hackroda
QA for the AI-built software era, built to help teams test code, workflows, and releases as they ship faster.
02 LiveKwanta
Retail commerce software that started as a client build: admin, POS, inventory, orders, and multi-channel sales.
03 BuildingNormata
Vertical software for repeated operating work in a focused service market, now being tested before deeper investment.
Jetisu platform
These create demand, operating context, and company support.
Public evidence, plainly stated.
Company economics and ownership remain in the data room. These are the operating facts we can state publicly as of July 2026.
15 Engineers and operators Across delivery and product
50+ Executives and team leads coached Through Devtx programs
3 Active product companies Hackroda, Kwanta, and Normata
2 Active operating markets United States and Kazakhstan
Jetisu network
The right people around a specific company.
Not passive community. A working group with a clear role.
See the networkFounders
Own the company and the problem.
Operators
Ship product and win customers.
Advisors
Bring domain judgment and access.
Angel investors
Fund specific companies and help them advance.
Current thinking
Notes from inside the work.
Field note: agents in production make cost an engineering problem
Running AI agents inside client systems taught us to treat tokens like any other metered resource.
Backing Hackroda: the bottleneck moved to verification
AI made writing code cheap. It did not make trusting code cheap. That gap is the company.
Services are the new software. We got there by accident.
Sequoia says the AI opportunity is labor spend, not tool spend. Our whole firm is that essay, lived backwards.
Work with Jetisu
Build, advise, or invest.
Bring a company, operating expertise, market access, or capital.