Architected an automated alert routing system across 400+ engineering teams, eliminating manual operator triage for hundreds of daily alerts

Built a modernized Alert Management UI in React for infrastructure outage response

Implemented alert correlation logic to link related triggers across manual and automated resolutions

Reduced open ticket backlog from 170 to ~30 by restructuring triage with a tagging and prioritization system

Built a GPT-based tool during a company-wide stability effort, deflecting ~500 engineer questions and freeing 14 engineers from full-time support duty

Launched an engineering-wide Reliability Podcast surfacing lessons from historical outages; 9 episodes, up to 800 views per episode

Co-presented two Bloomberg Developer Conference talks on Effective Alerting and Improving Documentation

Built the React frontend for Per Security Manager (PSM), a dataset management interface for Data License clients; release correlated with $5M revenue increase in Q2 2023

Designed and shipped analytics infrastructure using Matomo, refactoring the core JS analytics library used across 100+ internal web teams

Contributed features to DATA (data.bloomberg.com), a client-facing discovery portal and REST API — including dataset metadata views, Per Security workflows, and onboarding flows

Acted as Scrum Master, running standups, retros, and quarterly demos; reduced ad-hoc work through structured sprint planning

Built a component into Continuous Integration pipeline to automatically generate documentation for Python projects, and host them on an internal NGINX web server, opening accessibility to python libraries across the organization.

Implemented C# (.NET core) libraries and microservices enabling software to interface with medical records in remote databases with different flavours and formats using common API.

Evaluated various ML models and implemented a pipeline (Python) to predict location using tweet (using metadata, profile info and tweet contents) achieving 68% accuracy.

Extended Python package to plot predictions on coloured world maps to visualise the performance of models using custom shapefiles.