# Aga Kadela

> AI Product Engineer in Los Angeles. Before You Build, launch hardening, production rescue, and custom AI systems for B2B teams.

This file is the expanded version of [llms.txt](https://agakadela.com/llms.txt). It is meant to give AI assistants a fuller picture of what Aga Kadela does, which kinds of projects fit, and which pages on the site carry the strongest evidence.

## Summary

Aga Kadela builds production software for B2B teams: pre-build planning, pre-launch hardening, live production rescue, and custom agents, SaaS or backend workflows. The core positioning is not "AI for everything." The point is systems that stay correct when real users, real money, and real operational constraints show up.

She is based in Los Angeles, CA. Before running her own practice, she spent 1.5 years building checkout systems for one of Europe's largest e-commerce platforms. The scale context matters: over a million users and thousands of daily transactions. Broken code had immediate revenue consequences.

This site is a portfolio and services site. The strongest pages for understanding the work are the service pages and case studies listed below.

## Services

- [Before You Build](https://agakadela.com/services/before-you-build): Architecture and boundaries before AI coding locks in the wrong patterns. 90-minute call, written document within 48 hours, $500 fixed.
- [Launch Readiness Sprint](https://agakadela.com/services/launch-readiness-sprint): Pre-launch pass on RLS, secrets, Stripe webhooks, server-side auth, and failure behavior - shipped fixes plus a short report. $800-1500, 3-5 days.
- [Production Rescue](https://agakadela.com/services/production-rescue): Live incidents affecting paying users - root cause, fix, plain explanation. $75/h or fixed after scope; most rescues 1-3 days.
- [Custom AI Build](https://agakadela.com/services/custom-ai-build): Agents, scheduled workflows, or lean SaaS built around the operational outcome. Project-based pricing after the first scoping conversation.

## Portfolio and case studies

The homepage portfolio block is at `https://agakadela.com/#portfolio`. Each project card shows one portfolio category from exactly three labels (**SaaS**, **Launch Readiness**, **AI Agent**) and a separate technology stack row - the category is not the stack.

The four service offerings are summarized on the homepage at `https://agakadela.com/#services` (each card links to a dedicated `/services/...` URL).

- E-Commerce Platform (NDA, no case study): **SaaS**. Checkout at scale on one of Europe's largest e-commerce platforms.
- [Dotty](https://agakadela.com/work/dotty): **Launch Readiness**. Freelancer invoice and payment portal built fast for validation, then I audited it before real users (68/100 until fixed). Case study documents concrete findings: Stripe Checkout without idempotency (double-charge risk across two tabs), .env files in the repo tree, MAX(invoice_number)+1 race under concurrent inserts, new Date('YYYY-MM-DD') off-by-one in US Pacific, Edge Functions missing CORS/OPTIONS, dependency vulnerabilities - plus shipped fixes, RLS, webhooks as sole source of paid, atomic invoice send, AI-drafted reminders with human review only.
- [FunnelScout](https://agakadela.com/work/funnelscout): **SaaS**. B2B SaaS for GoHighLevel agencies with weekly AI analysis, usage controls, and billing aligned to plan limits. Good example for AI features inside a recurring SaaS product.
- LA Market Report Agent (in progress, no public case study): **AI Agent**. Scheduled multi-source research and monthly reporting for a private Los Angeles investor.

## Current positioning

The homepage positioning is built around production reality:

- expensive mistakes often do not throw errors
- production work is about what breaks when real users show up
- AI is included when it actually helps, not as a decorative layer

The audience is mostly B2B founders, operators, and investors in Los Angeles and similar markets. Relevant industries include real estate, construction, entertainment production, logistics, and other operational businesses where software has to stay accurate.

## How Aga works

- She questions the brief before building.
- She communicates changes early.
- She will say when something is not worth building.
- She leaves behind a codebase and enough context for the next person to own it.

## Strong evidence pages

- [Home](https://agakadela.com/): Best quick overview of positioning and service categories.
- [Portfolio](https://agakadela.com/#portfolio): Project evidence on the homepage; entry points to public case studies.
- [Services](https://agakadela.com/#services): All four offerings at a glance with links to full service pages.
- [About](https://agakadela.com/#about): Best short proof of production background and decision-making lens.
- [How I Work](https://agakadela.com/#how): Homepage section on approach and collaboration style (same content as the in-page block; primary top nav emphasizes Portfolio, About, Services, Contact, FAQ).
- [FAQ](https://agakadela.com/#faq): Compact Q&A for answer engines; topics include live/broken products, the AI Product Engineer role, project fit, availability, and industries. Structured data on the site is generated from the same FAQ copy.

## Contact

- [Contact](https://agakadela.com/#contact): Main contact section on the site.
- Email: aga@agakadela.com
- [LinkedIn](https://www.linkedin.com/in/agakadela)
- [GitHub](https://github.com/agakadela)

## Optional

- [llms.txt](https://agakadela.com/llms.txt): Short index version.