Hi,

I'm Riley, an accountant turned product leader.

tl;dr

  • Multiple roles building finance, contracts, and royalties platforms, from accounting and finance ops to product leadership at Amazon.
  • Led corporate reporting platforms across BlackLine, Workiva, and Oracle HFM, standardizing close for 1,600+ users, improved consolidation and close @ Amazon Financial Systems
  • Background: B.A. Accounting & Information Systems @ Gonzaga; started in accounting (Amazon Advertising, Financial Systems, Retial & Marketplace) then moved into product leadership @ Amazon Music
  • Dad mode, skis waxed, Boglehead nerd.

About Me

I'm an accountant turned product manager. I live in-between the CPA and the SDE and build finance and accounting systems that hold up with real money and real scrutiny.

Since 2017 at Amazon I've worn the controller hat, the program automation hat, and the enterprise product hat. I started in a four-rotation accounting program, owning close and reporting for businesses like IMDb and Amazon Advertising. Having lived in the dated accounting word, I've had first had experience with the manual journal entries, massive excel workbooks, and repetitive month-end close cycles. To capitalize on this opportunity I moved into a program manager role and scaled a self-service Alteryx automation program company-wide, shipping prebuilt integrations, training, and hundreds of compliant workflows that eliminated tens of thousands of manual hours. From there I led corporate reporting platforms across BlackLine, Workiva, and Oracle HFM, standardizing reconciliations and journals for 1,600+ users, accelerating SEC and statutory reporting, and cutting consolidation times by double digits.

Today I own Amazon Music's royalties stack (reporting, payments, and financial record keeping) where I've delivered a global platform managing billion-dollar obligations, lifted payment accuracy, eliminated penalties, re-architected the calc engine for minute-level runs, and shipped GenAI products that turn messy usage data into clear, auditable financial narratives. Through all of it, my edge is translating debits and credits into code that auditors trust and teams actually use

Riley Herges headshot

Experience

Sr. Product Manager

Amazon Music

Apr 2022 - Present

Senior PM at Amazon Music (Licensing and Reporting), owning the royalties, contracts, and financial record keeping platform, with strategy and delivery across Business Development, Accounting, Finance, Legal, and Engineering.

Sr. Product Manager

Amazon Corp Accounting

Jan 2020 - Mar 2022

Previously led corporate reporting platforms in Amazon Corp Accounting, owning BlackLine, Workiva, and Oracle HFM to standardize close and reporting across a large global user base.

Program Manager

Amazon Finance

2018 - 2020

Earlier, ran an accounting automation program, scaling self-service workflows and partnering closely with finance teams to remove manual steps and harden controls.

Finance Rotational Program

Amazon

2017 - 2018

Started in Amazon's Finance Rotational Program, including controller work for IMDb, enterprise recon administration, and support for large integrations like Whole Foods.

Other Project

Variance Insights

Plain-language close analytics. I built a product that turns raw usage and financial data into narrative explanations of month-over-month swings. It pairs governed drill-downs with readable summaries, so finance leaders can get from signal to story fast, then back to line-item detail when needed. The result was a step-function cut in analysis time, from a day to about an hour in typical cycles. It also reduced escalations, because the "why" traveled with the numbers. Adoption was driven by trust, not novelty, which is why we anchored on traceability and permissions from day one.

Variance Insights UI - Overview Variance Insights UI - Detail View
Analytics/BI, SQL, LLM Summarization,Finance FP&A

Close & Consolidation

Standard work at enterprise scale. Before the streaming domain, I led consolidation and close tooling for a large multi-entity environment. My focus was standard templates, stronger approvals, and automated reconciliations, all wired into the reporting flow users already lived in. We improved consolidation speed significantly and drove consistency across geographies and entities without sacrificing local needs. The value wasn't just faster close, it was fewer one-offs, cleaner audit evidence, and less stress on the last mile. I measured success by stability during peak load and by how often teams could self-serve vs. escalate.

Oracle, HFM, BlackLine, Close & Consolidation, SOX, GAAP, Change Management

Self-Service Automation

Freeing accountants from swivel-chair work. Earlier in my career I ran a self-service program that put governed automation in the hands of finance users. We shipped reusable patterns, integrations, and training so teams could retire manual steps while staying within policy. The program scaled globally, with hundreds of compliant workflows and tens of thousands of hours returned annually. The throughline was safe speed: codify the "right way," make it easy, and adoption follows. That playbook still shapes how I roll out internal products today.

Alteryx, Python, Low-Code Patterns, ETL Templates, Python/Scripts, Governance/Standards, Training/Enablement,

Just for Fun

Music Business News Digest

AI-generated news summarization platform built on AWS – self-initiated, widely used by colleagues

About:

In 2024, I launched a self-directed project to apply generative AI in a real-world use case: condensing the fragmented world of music business news into a daily digest tailored to licensing, royalties, and industry strategy. Inspired by internal requests for faster, curated news updates, I built and deployed an end-to-end summarization pipeline on AWS in under a week. The tool has since been leveraged by coworkers across Finance, Licensing, and Content Strategy teams to stay informed on relevant developments.

The system ingests headlines and article content from trusted music business sources (Billboard, Variety, Music Business Worldwide, etc.), summarizes each using Claude Sonnet 3.5 via Amazon Bedrock, and rewrites them into a polished newsletter-style markdown digest prioritized by internal interests (e.g., royalties, DSP strategy, publisher M&A). Summaries are stored and versioned in Amazon S3, and the full pipeline is orchestrated via Step Functions and Lambda, with support for concurrent processing and retry logic.

AWS Console showing the Music Business News Digest architecture Sample output from the Music Business News Digest

Key technical components:

  • Orchestration: AWS Step Functions map over ~10 curated URLs daily, with concurrency limits and per-article Lambda execution.
  • AI summarization: Claude Sonnet 3.5 (via Bedrock) generates article-level and digest-level summaries, with prioritization based on a dynamic "interest list."
  • Error handling: Each stage (URL fetch, summarization, store) supports retries and timeouts; malformed articles are logged to S3 for inspection.
  • Markdown generation: Final output is written in clean, human-readable markdown with clickable source links, optimized for internal distribution.
  • Frontend (WIP): A lightweight React + Cloudscape frontend allows users to trigger refreshes and browse historical digests from S3.

Challenges solved:

  • Bedrock throttling: Implemented backoff/retry strategy to gracefully handle ThrottlingException errors under high load.
  • Cold start latency: Optimized Lambda memory and prewarmed frequently used functions to reduce startup delays.
  • Input sanitization: Filtered junk DOM content and truncated article text to avoid exceeding token limits during summarization.
  • Digest clarity: Used Claude's structured output formatting to produce reliable, sectioned markdown summaries with linked sources.

Results:

  • 80% reduction in time spent manually scanning music industry news
  • Used consistently by internal teams across Finance, Royalties, and Licensing
  • Deployed and iterated entirely independently within 1 week
  • First personal proof-of-concept for building production-grade GenAI pipelines using Bedrock, Step Functions, and Claude
GenAI, Amazon Bedrock, Claude Sonnet, AWS Lambda, Step Functions, S3, Markdown, News Summarization, Royalty Strategy

Personal Finance Dashboard

React-based financial tracking application built with modern design libraries – fun side project to replace Excel-based personal finance tracker

About:

I built a comprehensive personal finance dashboard as a fun side project to replace my Excel-based financial tracker. The application transforms raw transaction and account balance data into actionable insights through interactive dashboards, budget tracking, and sophisticated data visualizations. Built entirely client-side with React and deployed as a static site, the tool provides a complete financial overview without compromising data privacy.

The system processes CSV exports from bank accounts and credit cards, categorizes transactions automatically, and generates monthly budget analyses with scoring algorithms. It tracks net worth progression across asset categories (cash, investments, retirement, 529 college savings) and provides detailed spending breakdowns by category. The crown jewel is an interactive Sankey diagram that visualizes monthly income flow from gross earnings through deductions, retirement contributions, savings, and spending categories.

Personal Finance Dashboard - Financial Overview Personal Finance Dashboard - Multi-Month Budget Personal Finance Dashboard - Income & Flows Sankey Personal Finance Dashboard - Net Worth Account Balances

Key technical components:

  • Frontend Architecture: React 18 with Ant Design component library for consistent UI/UX, featuring responsive design and dark/light theme switching.
  • Data Visualization: Plotly.js for interactive Sankey diagrams, Recharts for time-series charts and pie charts, with custom color-coded flows for income, deductions, retirement, and spending categories.
  • Data Processing: Client-side CSV parsing and categorization engine with intelligent account type inference (checking, savings, 401k, etc.) and transaction categorization mapping.
  • State Management: React hooks with localStorage persistence for budget configurations, income settings, and user preferences across browser sessions.
  • Budget Intelligence: Custom scoring algorithm that compares actual spending against budget targets with weighted category analysis and month-over-month trend tracking.
  • Development Tools: Built with Cursor as primary IDE and deployed via Vercel for quick iterations and seamless static site hosting.

Challenges solved:

  • Data Integration: Built flexible CSV import system that handles various bank export formats and automatically categorizes transactions based on description patterns and custom mapping rules.
  • Performance: Optimized client-side processing for large transaction datasets with efficient data structures and lazy loading for visualization components.
  • Data Privacy: Designed entirely client-side architecture ensuring no financial data ever leaves the user's device, with localStorage persistence for user convenience.

Results:

  • Replaced complex Excel-based tracking with streamlined React interface
  • Automated transaction categorization reduces manual data entry by 90%
  • Interactive Sankey diagrams provide clear visualization of income flow patterns
  • Real-time budget scoring with month-over-month trend analysis
  • Complete client-side solution maintains full data privacy
React 18, Ant Design, Plotly.js, Recharts, CSV Processing, Data Visualization, Client-Side Architecture, Budget Intelligence, Financial Analytics