Experience

My professional journey so far.

  1. Software Engineer Intern

    Plexusphere Labs

    Jan 2026 — Present · Seattle, WA

    • Contributed to and optimized components of an AI Agent pipeline for transforming long‑form narratives into structured screenplay elements, character extraction, scene segmentation, and plot analysis, enabling downstream generation workflows
    • Built a scalable AI image generation system on text‑to‑image APIs with dynamic model selection, combining prompt engineering with a reference image library to ensure consistent character representation across outputs
    • Contributed to a model‑agnostic storyboard‑to‑video pipeline by transforming storyboards into structured scene data, implementing automated shot sequencing and video composition, and developing a React UI for scene editing, preview, and generation
    • Designed and shipped Orbit, a full‑stack content operations platform for creators managing multiple content verticals, featuring multi‑workspace Kanban, a drag‑to‑schedule calendar, multi‑platform publishing coordination, and a performance analytics dashboard
    • Integrated AI‑powered workflows into Orbit, RSS‑driven topic discovery, cached daily/weekly AI reports, and context‑aware brief generation (audience profile, hook analysis, title candidates, outline), through an OpenAI‑compatible API layer with configurable settings
    • Engineered the Orbit frontend with React, TypeScript, Vite, and Tailwind CSS; managed state with Zustand and TanStack Query; localized to 5 languages via react‑i18next; and shipped as an offline‑capable PWA using Workbox and Dexie.js
    • Built the Orbit backend with Fastify (Node.js, TypeScript) on PostgreSQL, with Redis for cache/sessions and BullMQ for asynchronous job processing, secured by JWT + OAuth 2.0 (Google), and deployed via Docker Compose

  2. Data Science Intern

    Chongqing Fanyue Logistics Co., Ltd.

    May 2025 — Aug 2025 · Remote

    • Developed predictive models using Random Forest and XGBoost to forecast package delivery times, improving on‑time delivery rates by 12%
    • Analyzed over 2 million shipment records using Python and SQL to identify and address bottlenecks in regional distribution centers
    • Designed interactive dashboards in Tableau and Matplotlib visualizing real‑time KPIs (volume, on‑time rate, cost per shipment, hub utilization) for operations managers, replacing manual Excel reports
    • Conducted A/B testing for various routing strategies, extracting actionable insights that led to a significant 10% reduction in average delivery costs
    • Utilized regression analysis to accurately project future expenses, directly contributing to a 3% increase in overall budget forecasting accuracy
    • Collaborated cross‑functionally with data engineers and operations teams to deploy automated data pipelines for daily shipment tracking reports