App Builder

Consumer software shaped by workflow reality.

I build AI-assisted apps for personal records, receipts, and evidence. The work combines product judgment, secure cloud backends, extraction quality, and supportable user workflows.

Featured product

ExpenseJournal

MVP / QA

ExpenseJournal is a personal and household receipt evidence assistant for receipt capture, OCR/AI extraction, uncertain-field review, line-item and semantic search, and confirmed evidence reports.

01Capture

Camera, upload, or forwarding for receipts and evidence.

02Extract

OCR and AI-assisted merchant, total, tax, and line-item extraction.

03Review

Focus the user on uncertain fields before records become trusted.

04Search

Find records by item, vendor, amount, date, category, or meaning.

05Export/Report

Create confirmed evidence reports for taxes, warranty, reimbursements, or household review.

Product media

ExpenseJournal product screens

Current product screens showing the receipt capture, review, search, reporting, export, and household-sharing workflow.

Screens are product/marketing previews from the current ExpenseJournal build.

Capabilities

What ExpenseJournal is built to handle.

Receipt capture

Mobile-friendly capture and upload paths for everyday purchasing records.

OCR/AI extraction

Structured extraction with human confirmation before evidence becomes trusted.

Line-item search

Search by product, vendor, category, date, amount, and related meaning.

Confirmed reports

Evidence packages for warranties, reimbursements, tax preparation, and household review.

Household sharing

Shared records and practical ownership models for more than one user.

Operational telemetry

Quality signals for extraction failures, review patterns, and supportable production behavior.

Shared product architecture foundation

Reusable platform foundation.

This is the reusable platform behind ExpenseJournal and future personal records products: secure identity, document processing, extraction quality, semantic search, trusted reporting, governance, and operating telemetry.

01Capture

Receipts, documents, images, and user-provided records.

02Extract

OCR and AI-assisted structured field and line-item extraction.

03Confirm

Human review where confidence, policy, or evidence matters.

04Search

Line-item, field, date, vendor, and semantic retrieval patterns.

05Report

Confirmed evidence packages and practical summaries.

06Govern

Privacy, usage limits, data boundaries, and ownership.

07Observe

Telemetry for extraction quality, failures, review loops, and support.

Next direction

Personal health organizer

In development

A privacy-focused direction for organizing user-provided routines, symptoms, habits, reminders, documents, and health notes so personal information is easier to find and summarize.

Not intended for diagnosis, treatment, or medical advice.

Reusable foundation

Shared cloud and AI platform

  • Identity and privacy-aware access patterns
  • Document/image processing and OCR
  • Extraction, semantic search, and secure storage
  • Telemetry, quality signals, and support workflows
  • Usage controls and subscription-aware limits

Applied AI patterns

Applied AI patterns behind the product.

The product work is intentionally grounded in user-visible evidence, review, and support paths instead of vague automation claims.

Extract with confidence

OCR/AI extraction, extraction confidence, uncertain-field review, and human-in-the-loop confirmation before records become trusted.

Retrieve with evidence

Semantic search, retrieval, evidence-grounded summaries, line-item references, and confirmed reports that users can inspect.

Operate with signals

Telemetry and quality signals for extraction failure, review loops, support paths, cost/latency awareness, and fallback behavior.

Respect the boundary

Privacy boundaries, usage controls, secure storage, practical ownership, and supportable workflows for personal records.

What this demonstrates

Product work with architecture discipline.

Product judgment

Choosing workflows where AI helps because the evidence and review path are concrete.

Architecture discipline

Designing storage, identity, telemetry, and ownership before the product surface grows.

AI workflow design

Keeping extraction, confidence, review, and search grounded in user-visible records.

Production readiness

Building for support, auditability, rollback paths, and measurable quality signals.