Tom Moroney

Tom Moroney

Software Engineer at Workday

I'm a software engineer from Dublin, currently working full-time at Workday on backend security and developer tooling for enterprise systems, including sensitive-data redaction, controlled access workflows, and AI-assisted documentation.

Outside of work, I build AutoSubs, an open-source, cross-platform desktop app helping hundreds of thousands of creators automatically generate subtitles for their videos in any language, saving them time while making their videos more engaging and accessible to global audiences. I'm also developing SubSlate, a new kind of video editor.

If you wish to get in contact, drop me an email at tom.moroney.dev@gmail.com.

Experience

Software Engineer at Workday
  • Architected a high-performance .NET and Java/Log4j redaction framework that removes sensitive customer data from logs and obfuscates exception stack traces across 10 microservices with minimal performance impact.
  • Coordinated rollout and testing with service owners, contributing to Adaptive Planning's FedRAMP audit preparation for federal approval.
  • Built a secure Java lookup API, integrated with the .NET lookup tool UI, letting support teams map internal IDs to sensitive data and run reverse lookups through controlled access paths.
  • Developed an AI-powered living documentation and Q&A system that summarises code from AST leaf nodes upward, keeping docs synchronised and answering developer queries.
Software Engineer Intern at Workday
  • Engineered 4 Java/Spring Boot APIs for database bootstrapping and multi-tenant initialisation, supporting Adaptive Planning's migration to Workday's unified cloud platform.
  • Implemented transactional rollback, SQL fixes for import defects, and a database-simulating test framework achieving 92% line coverage, improving tenant migration reliability.
Software Engineer Intern at ChannelSight
  • Developed web crawlers for ecommerce product and review data, analyzing DOM structures and network traffic to create resilient selectors and extraction logic.
  • Implemented proxy rotation, custom request headers, and rate limiting to reduce scraper failures from layout changes and bot-detection mechanisms.
Class Representative at Trinity College Dublin
  • Elected as class representative for around 60 Integrated Computer Science students, acting as a liaison between students and faculty to address concerns and improve the academic experience.

Education

Trinity College Dublin

Masters in Computer Science (M.C.S.)

First Class Honours

Trinity College Dublin

B.A. (Moderatorship) in Computer Science

Second Class Honours

Featured Projects

AutoSubs

Open-source, cross-platform desktop app for video creators to generate, translate, label speakers with unique per-speaker styling, and place subtitles directly into video editor timelines. Built with React/TypeScript, Rust, WebView, LuaJIT, and custom Adobe extensions.

  • Built and maintained over 2+ years with 3.5K+ stars and 400K+ downloads, using a Rust backend to stay under 200MB idle alongside video editors.
  • Developed editor integrations for DaVinci Resolve via an embedded LuaJIT HTTP bridge and Adobe apps via custom extensions for timeline audio extraction and subtitle placement.
  • Implemented audio preprocessing and on-device model management, automatic language-based model selection, translation, and a formatting engine optimized for CJK, Korean, RTL, Indic, and SE Asian line-breaking and timing constraints.
  • Rust
  • TypeScript
  • React
  • Lua
  • 3.5K+ stars
  • 400K+ downloads

Beyond Automation: AI Framework for Intelligent Cinematographic Editing

Multimodal video-editing framework that reached 72.5% shot attribute classification accuracy, fusing visual composition, optical flow, audio-temporal patterns, and narrative reasoning for next-shot recommendations.

  • Achieved 72.5% accuracy in shot attribute classification, a 16 percentage point improvement over the Anatomy of Video Editing benchmark.
  • Developed a multimodal pipeline optimized for Apple's MLX framework, enabling real-time, on-device inference by fusing visual composition, optical flow, and audio-temporal patterns.
  • Implemented a narrative reasoning engine for next-shot recommendations using VLMs and LLMs to bridge technical pattern recognition and semantic story comprehension.
  • PyTorch
  • MLX
  • VLMs
  • LLMs
  • Video AI