AI / data / integrations
ListBackup.ai
A data backup and integration platform spanning connected services, serverless workflows, and operational safeguards.
Selected work
Three platforms, described from the code rather than the pitch. The repositories are private, so what is on this page comes from the projects’ own README and architecture documents and stops where the client’s permission stops. No invented numbers, no logos we have no right to show.
AI / data / integrations
A data backup and integration platform spanning connected services, serverless workflows, and operational safeguards.
CRM / automation
A CRM automation platform connecting helpers, workflows, and integrations for teams that run on their customer data.
Consumer AI / platform
Current platform and AI work, shared only at the public, high-level product direction the client has published.
A platform that pulls data out of the services a business runs on, keeps versioned copies, and moves that data where it is needed next.
Product and platform engineering
AWS serverless: Go Lambda services, DynamoDB, S3, Cognito
Next.js 15, React 19, TypeScript, Tailwind, Radix
A business keeps its real records inside other people’s software. The CRM, the mail platform, the billing system. Each of those has an export button and none of them has a memory. When a sync goes wrong or somebody deletes the wrong list, there is nothing to go back to.
Connect to fifteen or more platforms, back them up on a schedule or on demand, and keep the copies in per-account S3 storage with versioning, and point-in-time recovery on the DynamoDB metadata. Metadata and activity logs live in DynamoDB so every run can be inspected.
Keap, Stripe, GoHighLevel, ActiveCampaign, HubSpot, QuickBooks, MailChimp, Zendesk, Google Drive and more, plus REST for anything else. Each one gets an OAuth flow, and the tokens are encrypted before storage, with client credentials and encryption keys held in AWS Secrets Manager.
Every account gets its own S3 bucket and its own encryption keys, and accounts nest without a fixed depth limit. That is what lets one platform serve a sole trader and a group of companies without separate deployments.
Roughly twenty-nine modular Go Lambda services sit behind one HTTP API with a custom JWT authorizer. The Node.js tree is kept as legacy reference and is not deployed. A deployment catalog records each service's prerequisites so the order is checked by the pipeline rather than remembered from a runbook.
Everything above comes from the project’s own README and ARCHITECTURE.md. Those repositories are private, so this page stays inside what those documents say. The platform’s own customer names, data volumes, revenue, and the decisions that are not written down stay off it entirely. On a call we can go through the parts that matter to your problem, including what we would do differently a second time.
How we talk about work
Follow the links to live products, published documentation, and our public profile to see the work for yourself.
If it is a client's number, we do not print it. No uptime percentages, no revenue lifts, no traffic charts you cannot check.
If we only advised, we say advised. Roles on this page describe what we actually did, not what the team around us shipped.
If you want the detail, ask on the call. Most of what is useful about a project cannot be published, and we will tell you why.