Software Development Engineer 2
Operational Reports 2.0: issuers, acquirers and affiliates use it to find, subscribe to and download 400+ payment reports. 20,000 users worldwide.
- Own the frontend for the whole app (report search, chatbot, recommendations, subscriptions, exports), working with product and business owners as requirements kept changing. I also handle production incidents, tracing them through Splunk logs.
- Built the frontend and the report-matching logic for the app's production RAG chatbot. Answers that mention a report show a different card depending on whether the user is subscribed to it. I also took part in the UI/UX discussions on the generative UI.
- Prototyped an on-device search agent on Chrome's built-in LLM, with no server round trip. Chrome has no native function calling, so a JSON-Schema-constrained intent extractor turns natural language into a typed query, calls the real backend API and feeds the results back for a summary. It also exposes a WebMCP tool schema.
- Built and demoed internal LLM agents (a defect-solver bot, a feature-implementation bot and automated testing agents) that got the team using AI tools in its development workflow.
- Owned the microservices, frontend and ETL pipeline work for the modernization initiative that moves the platform's services to microservices on AWS under PCI-DSS, where nothing can be stored unencrypted at rest, while keeping live services running.
- Led the rewrite from Angular to React, cutting load time by 75% (Largest Contentful Paint from 6s to 1.5s). Split the Billing Query Tool out of the monolith into its own micro-frontend with its own Jenkins CI/CD pipelines, and owned its release to production. 500+ users.
- Measured a 69% to 81% drop in logging tail latency (p99 and p99.9 at 8 threads) with a JMH benchmark, then wrote a hybrid design backed by a crash-loss simulation and a phased canary rollout. It was adopted into follow-up work.