Work

  • BloombergSoftware Engineering InternPythonC++ReactSQLBackend services & APIsFull-stack developmentGraph data modelingFixed-income analytics

    I built a service that maps how calculations depend on each other inside Yield and Spread (YAS) calculator, the Bloomberg Terminal’s bond analytics calculator. YAS is a top-10 Terminal feature with 300K+ daily hits, and the service models 15+ critical field dependencies in a system that handles 1M+ API requests a day. I also built a full-stack web app on top of it so engineers can store, search, and visualize those dependency graphs. That makes tracing and debugging fixed-income calculations a lot easier.

  • DatadogSoftware Engineering InternGoReactTypeScriptSnowflakeMCPSSO / RBACMetabaseAWS S3Slack integrations

    I built an AI chatbot for the revenue engineering team. I took it from scoping sessions with 3 product managers through to a secure Snowflake MCP server in Go, with SSO and role-based access controls, plus a backend agent service for AI-powered data investigation. I also added Snowflake dataset support to Datadog’s internal data catalog across the Go backend and React frontend, which let 8,700+ Snowflake users catalog and manage 8,400+ datasets. On top of that, I built an MCP server that lets coding agents like Claude Code and Cursor search 10,000+ internal Metabase dashboards and questions. I also rolled out Slack integrations to 35 users across teams and automated weekly 100MB+ data exports for the marketing team.

  • J.P. Morgan ChaseSoftware Engineering InternPythonPySparkDatabricksApache AirflowAWS MWAAAWS LambdaAWS S3Terraform

    I built distributed data pipelines for an on-demand analytics tool that measures structural interest rate risk across the bank’s assets and liabilities. The pipelines ran in Airflow on AWS and fed a real-time risk modeling platform in Databricks. I used Lambda to trigger them on demand, set up the infrastructure with Terraform, and connected decision models so they ran directly on PySpark clusters. At the end of the summer, I presented the architecture to 70+ engineers, managers, and interns.

  • VT Hume CenterResearch InternRR ShinyBayesian statisticsBeta-Binomial modelsData visualizationReliability analysis

    I built a browser-based R Shiny app that automates Bayesian reliability analysis for defense system testing. It uses a Beta-Binomial model to turn test results into full posterior distributions and checks them against 90–99% reliability thresholds. With it, 10+ researchers get results and visualizations in under 5 minutes instead of writing R scripts by hand. I wrote an 8-page technical report on the work and presented it at the 2025 Hume Center & IC CAE Colloquium and the Astronaut Scholar Foundation's Symposium, advised by Dr. Justin Krometis.

  • American Electric PowerEnergy Delivery InternPLS-CADDLiDARFinite element analysisQA/QCTransmission line modeling

    I modeled transmission and distribution lines from millions of LiDAR survey points in PLS-CADD to analyze clearances. I also ran finite element analysis and QA/QC checks to confirm projects could actually be built.