Portfolio

Four ventures. One standard of proof.

Each venture serves a different part of ordinary life. None of them is allowed to skip a gate. This page is the honest state of the portfolio: what exists, what it demonstrates, and what it still has to earn.

Ventures 4 Private QA 1 Pilot preparation 1 Research 2
Private QA

PoseLane

Creative collaboration

A working product for showing creative work, starting from a private brief, and moving toward an agreement without leaving the platform.

The human need

Creative work is discovered through feeds that flatten it. Turning interest into a brief, a scope, and an agreement usually means three tools, two inboxes and a lot of trust with no record.

What we built

Portfolio and discovery surfaces, a private brief flow, a request that becomes an agreement, and a consent-led assistant review that sits inside the flow rather than beside it.

Where it stands

The build is in isolated QA. The surfaces work end to end with demo data. No public user has used it, and no transaction has been recorded.

PoseLane QA capture: discovery surface with search and filters
Step 01

Discovery

Search, category filters and result cards on the discovery surface.

Capture from build v01.00.00.1909 · demo data. Not a live demo.

Three captures, one build. The tour changes which surface you are looking at; it does not simulate the product.

Working product

EatCue

Everyday food decisions

A mobile decision layer for food: browse, save, compare, and plan a short trip around what actually matters to you right now.

The human need

Choosing where to eat is frequent, low-stakes and strangely expensive in time. Reviews are noisy, distance is ignored, and the decision restarts every day.

What we built

A gallery and ordering flow, distance-aware discovery, and a save-and-compare layer that keeps a decision warm between sessions.

Where it stands

The build runs on device. The catalog behind it is not yet verified, and no pilot cohort has been measured, so usage and retention remain open questions.

EatCue QA capture: mobile food gallery
EatCue · 0.0.9+24 · demo data, captured on device
Research

CareTrip

Non-clinical care coordination

Coordination and trust for the non-clinical parts of care: appointments, transport, and the hand-offs between family members.

The human need

Families coordinate care with screenshots, phone calls and memory. The clinical record exists; the logistics around it usually do not.

What we built

Nothing yet. This venture has not passed gate 00.

Next to prove

Problem interviews with families and coordinators, a map of the non-clinical journey, and a consent model that can be trusted before a single screen is built.

Why it is public in this state: a portfolio that only shows finished work is a brochure. Research-stage ventures are listed so the cost of the next gate is visible.

Research

StudyBase

Tutoring and supplementary learning

A shared view of what a student actually needs, with AI assistance scoped carefully enough to be useful to a tutor.

The human need

Parents, tutors and students each hold a different picture of the same gap. Supplementary lessons fill time without a shared record of what changed.

What we built

Nothing yet. This venture has not passed gate 00.

Next to prove

A tutor and student workflow map, a written scope for AI assistance with safety review, and a pilot design with a single tutoring cohort.

No AI claim without a guardrail. If assistance cannot be explained to a tutor and a parent, it does not go into the build.

The portfolio is one system, not four bets.

Identity, trust, reputation, matching, booking, payment, dispute handling and an assistant layer are designed once and reused. Each venture that passes a gate makes the next build cheaper, which is the only compounding this model claims.