Skip to content

Software EngineerSan Francisco, CA

Building Distributed Systems, Scalable Data Pipelines & Agentic Infrastructure

View resumeGet in touch

About

I'm a software engineer at Capital One, working on the Data Discovery platform — the search, recommendations, and pipelines that help people across the company find the datasets they actually need.

Most of my work sits at the seam between data infrastructure and product: PySpark ETLs feeding a property graph in NeptuneDB, an OpenSearch KNN pipeline behind dataset recommendations, a Go and Kafka service handling 15K+ TPS — and the React frontends that surface all of it.

I studied CS at Berkeley. Lately I've been building agentic AI tooling to automate the infrastructure work I'd rather not do by hand.

Languages
PythonTypeScriptGoJavaSQL
Data
SparkKafkaSnowflakeDatabricksOpenSearchRedisPandas
Infrastructure
AWS (ECS, Lambda, EMR, DynamoDB, Neptune, S3)CloudFormationDocker
Frontend
ReactLitNext.js

Experience

Capital One

San Francisco, CA

  1. Software Engineer

    Nov 2025 — Present

    I own the ETLs and APIs behind the Data Discovery homepage, plus the frontend that surfaces them — the goal being one place where any data user can find, understand, and manage their assets.

    • Built React and Lit frontend components for the Data Discovery homepage — dataset recommendations, FTUX flows, data management tools, and notifications — with client-side state, a data layer, API integration, and event handling.
    • Overhauled legacy vector search infrastructure with an optimized OpenSearch KNN pipeline for dataset recommendations, improving hit rate by ~50% and reducing runtime by 63%.
    • Accelerated developer efficiency by 90% by building custom agentic AI skills that autonomously handle infrastructure provisioning, repo synchronization, and vulnerability patching.
    • Designed a property graph model for users, assets, and interactions, and built 12 PySpark ETL processors on EMR ingesting 50M+ records into NeptuneDB.
    • Architected an ML training dataset schema and pipeline on Databricks and Snowflake with retry logic, multi-asset extensibility, and end-to-end observability.

  2. Associate Software Engineer

    Feb 2024 — Nov 2025

    Took dataset recommendations from zero to one, then helped replace a third-party payments vendor with an in-house transaction engine.

    • Built an algorithms-driven dataset recommendations pipeline 0→1 for an internal data discovery platform, optimizing user-dataset ranking via extensive offline and online trials and feature weight tuning.
    • Developed an event-driven Go-based Kafka consumer and DynamoDB upsert handler achieving 15K+ TPS, enabling deprecation of a third-party payment vendor and slashing infrastructure costs.
    • Led an org-wide migration of secrets management from HashiCorp Vault to AWS Secrets Manager, completing cutovers for high-traffic ECS and Lambda services.

  3. Software Engineer Intern

    Jun 2023 — Aug 2023

    • Reduced cross-team coordination by ~1 week per sprint by designing and iteratively decoupling RBAC permission logic from upstream services within a new workflow management platform.
    • Led onboarding of the platform's first clients by transforming custom workflows into configurable setups.

Anaplan

San Francisco, CA

  1. Software Engineer Intern

    Jun 2022 — Aug 2022

    • Integrated an in-memory cache to throttle user traffic across API instances and optimize data retrieval times by 80% for downstream APIs with Redis and Java Spring Boot.
    • Designed exhaustive tests for a rate limit filter using asynchronous programming, multi-threading, and locks.

Projects

Tooling · 2025

Resume tailoring pipeline

A local pipeline that takes a job description and produces an ATS-clean, one-page resume from a single master context document, compiling LaTeX to PDF and logging every application. Built because tailoring resumes by hand does not scale.

  • Python
  • LaTeX
  • Claude Code

Placeholder · 2026

Your next project

Replace this entry in src/content/projects.ts. Keep descriptions concrete — what the thing does, the interesting constraint, and what you'd do differently. Two or three sentences is plenty.

  • TypeScript
  • Postgres

Education

University of California, Berkeley

2023

B.A. Computer Science3.9 major GPA

Coursework

  • Operating Systems
  • Database Systems
  • Computer Security
  • Algorithms
  • Machine Structures