Full-stack engineer

Contents

Each one covers what I chose, what I turned down and what I left alone.

1. Chapter 01 Stopping bad book data before it reached retailers Records with missing or contradictory metadata can no longer be switched on for retail feeds. At rollout, the new rule engine found 3,507 formats already in the feed failing at least one blocker, across 177 publisher accounts. 2 min read
2. Chapter 02 Fixing royalty statements that paid authors the wrong amount Author royalty statements were wrong in several independent ways, overpaying some authors and underpaying others. One opening statement showed $66.18; the author was owed $3,277.88. 2 min read
3. Chapter 03 Migrating a publisher and $5.5M of sales history A publisher’s royalty history came across onto their existing catalog without duplicating it. Of 451 MetaComet projects, 376 matched and the other 75 went to a reconciliation CSV, and 135,005 rows of sales history ($5.53M) came across. 3 min read

Earlier work

Smaller wins

• An upload race in feed delivery. A race in a Lambda’s S3 upload broke a 1,258-product ONIX feed. I fixed it by switching from a stream upload to a buffer upload.

• FTP delivery. FTP and SFTP shared one code path, and that client only speaks SFTP, so every FTP partner’s delivery failed. I gave FTP its own client and split the path.

• Retail file names. A file-naming system for book assets sent to retail partners: files are renamed to the ISBN-based naming convention retail partners expect.

Index

The stack I work in, and the ideas behind the chapters. Numbers are chapters and link to them; E is earlier work, and W a smaller win.

• Amazon S3, 3: Chapter 3: Migrating a publisher and $5.5M of sales history, W: Smaller wins: An upload race in feed delivery
• AWS Lambda, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount, W: Smaller wins: An upload race in feed delivery
• Backfills, 1: Chapter 1: Stopping bad book data before it reached retailers, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• C3 AI APIs, E: Earlier work: Transformer Alerts
• C3 AI Suite, E: Earlier work: Transformer Alerts
• CSV import adapters, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Data migration, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Diagnostic scripts, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• Discount bands, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• Express, 1: Chapter 1: Stopping bad book data before it reached retailers
• fast-xml-parser, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Fuzzy name matching, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• ISBN-13 checksums, 1: Chapter 1: Stopping bad book data before it reached retailers
• JavaScript, E: Earlier work: Transformer Alerts
• Jest, 1: Chapter 1: Stopping bad book data before it reached retailers, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• Legacy royalty data, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• MySQL, 1: Chapter 1: Stopping bad book data before it reached retailers, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Next.js, 1: Chapter 1: Stopping bad book data before it reached retailers
• Node.js, 1: Chapter 1: Stopping bad book data before it reached retailers, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• ONIX 3.0, 1: Chapter 1: Stopping bad book data before it reached retailers, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Pug, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• React, 1: Chapter 1: Stopping bad book data before it reached retailers, E: Earlier work: Transformer Alerts
• react-hook-form, 1: Chapter 1: Stopping bad book data before it reached retailers
• Reconciliation, 3: Chapter 3: Migrating a publisher and $5.5M of sales history
• Royalty statements, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• Sequelize, 1: Chapter 1: Stopping bad book data before it reached retailers
• SFTP/FTP, W: Smaller wins: FTP delivery
• Statement periods, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount
• Transactions, 1: Chapter 1: Stopping bad book data before it reached retailers
• TypeScript, 1: Chapter 1: Stopping bad book data before it reached retailers, 2: Chapter 2: Fixing royalty statements that paid authors the wrong amount, 3: Chapter 3: Migrating a publisher and $5.5M of sales history, E: Earlier work: Transformer Alerts
• Validation rules, 1: Chapter 1: Stopping bad book data before it reached retailers
• zod, 1: Chapter 1: Stopping bad book data before it reached retailers

Also worked with: AWS RDS, Puppeteer, Radix UI, Slack API, Stripe, Tailwind, TanStack Query, TanStack Table, XSD.