Work so far

3 years as a fullstack software engineer , highlights below.

  1. Fastr

    Software Engineer — Fastr

    Boston, MA · 2023–2024

    Platform reliability + developer tooling at startup pace. Extended the design editor and wired up observability + CI/CD so teams could ship faster with confidence.

    Adoption

    40+ e-commerce customers

    Release cadence

    +25% faster

    Uptime

    -40% outages

    • Extended the Penpot-based editor with ADA compliance & responsive layouts (Clojure + TypeScript).
    • Shipped analytics dashboards (ClojureScript UI + gRPC/SQL backend) → 35% faster insights and -500 ms load time.
    • Automated build/test/deploy with GitHub Actions/Jenkins + Docker to accelerate releases 25%.
    • Integrated Grafana + Kibana → -20% time-to-detect and -15% error rate.
    • Closed 30+ production issues within 24h SLAs; mentored 2 juniors to speed up onboarding.
    ClojureTypeScript PythongRPC PostgreSQLDocker GitHub ActionsGrafanaKibana
  2. Cvent

    Software Engineer — Cvent

    2022–2023

    Full-stack on Attendee Hub: cleaner UIs, faster APIs, and tighter delivery loops across a large user base.

    UX complaints

    -30%

    API latency

    -20%

    Sprint velocity

    +15%

    • Built planner & guest-facing features with React/TypeScript/Next.js → -30% navigation complaints.
    • Implemented 4 backend services in Java + Dropwizard used by 50k+ event users.
    • Connected Apollo GraphQL to REST backends; dashboards saw 20% faster responses.
    • Improved CI (Docker, test strategy, parallelization) → +15% sprint velocity and fewer rollbacks.
    • Designed caching & indexing for event workflows → -40% query time; documentation cut escalations -25%.
    JavaDropwizard ReactTypeScriptNext.js Apollo GraphQLPostgreSQL DockerJenkins

Hands-on Projects

Scalable Financial Analytics Pipeline

Built a distributed analytics pipeline using Hadoop, Spark, and Docker for processing financial time-series data. Implemented anomaly detection, clustering, and volatility prediction in a scalable system.

HadoopSpark PySparkDocker

Symbolic Music Generation

Built a neural network pipeline using Python and Keras to generate piano music from MIDI datasets using sequence modeling techniques.

PythonKeras Deep Learning

Land Use Classification

Classified urban land use using satellite imagery and machine learning models like Random Forest and SVM in Google Earth Engine.

Google Earth EngineML Remote Sensing

Recipe Finder Web App

Developed a React-based SPA to search recipes via REST APIs with pagination and efficient state management using React Hooks.

ReactJavaScript REST API