CV

Software engineer working on backend and geospatial data platforms, with a background in large-scale data infrastructure.

Download résumé (PDF) · GitHub · LinkedIn · Google Scholar

Experience

Software Engineer, Dyneti Technologies — San Mateo, CA · Jun 2026 – Jul 2026

• Designed, developed, tested, and deployed an end-to-end asynchronous processing workflow across a Java/Kotlin Android SDK, Node.js backend, and AWS RDS, cutting user-facing latency by over 75% by decoupling server-side ML inference from the synchronous request path
• Built a server-to-server decision API allowing partner backends to retrieve results precomputed at request time, eliminating a redundant second call and query-time inference delay
• Defined the async lifecycle across mobile and backend: API contracts, RDS-backed schemas, inference state transitions, and pending-result handling, while preserving backward compatibility for live partner integrations
• Owned releases end to end via GitHub Actions CI/CD, pairing automated unit/integration tests with multi-device instrumented tests covering async workflows, failure handling, and edge cases

Software Engineer Intern, IBSS Corp — Remote, US · Sep 2025 – May 2026

• Built TerraPrecise, a Python/FastAPI data platform ingesting forecast, station, and terrain datasets, storing gridded outputs in S3, and serving location-specific weather risk products to React/Leaflet clients and automated alerting workflows
• Implemented the data layer around access patterns: PostgreSQL/PostGIS for subscriber, location, and forecast metadata; TimescaleDB for location-level forecast and observation time series; and S3 for large gridded outputs
• Designed a cloud-native geospatial serving architecture using eoAPI, STAC, COG, and TiTiler, making gridded weather datasets discoverable, queryable, and servable as scalable map layers through standardized APIs
• Designed and deployed a Dockerized microservice architecture on AWS isolating GPU-based downscaling model inference from lightweight API and alerting services through shared data contracts, with Lambda supporting event-driven ingestion
• Built reproducible processing and validation workflows integrating HRRR forecasts, station observations, terrain, and high-resolution model outputs, with automated quality evaluation against independent ground observations
• Shipped an internal Vertex AI chatbot streamlining HR support workflows, automating its knowledge-base updates with scheduled Cloud Run jobs syncing documents from Google Drive to Cloud Storage

Software Engineer Intern, TGS — Houston, TX · May 2024 – Aug 2024

• Built parallel Python data pipelines (Xarray/Dask) on GCP processing 20 TB of numerical weather predictions and offshore lidar observations for a wind resource assessment platform
• Developed data-quality and ML-based bias-correction workflows combining model forecasts with observational data, improving wind-speed RMSE by 15% and producing calibrated datasets for downstream analytics
• Cut data processing time 70% by redesigning Zarr and Parquet storage layouts and chunking strategies, letting analytics jobs handle larger datasets within fixed memory limits

PhD Researcher (funded by NASA), University of Virginia — Charlottesville, VA · Aug 2019 – May 2025

• Wrote reusable parallel Python pipelines (Xarray/Dask) for ingestion, quality control, and feature extraction across 100+ TB of satellite, reanalysis, and model-derived geospatial data in Unix/Linux environments
• Built a lightweight HTML/JS/Leaflet interface to visualize spatiotemporal analysis results from NetCDF datasets, enabling interactive exploration across time and location

Projects

See Projects for detail.

• Synchronized Distributed File System — gRPC-based DFS in C++ with whole-file caching, CRC32 diffing, timestamp-based conflict resolution, server-side writer locks, and mutex-guarded concurrent client sessions
• Multi-Agent Travel Planner — eight Google ADK agents composed into a sequential pipeline with a bounded critic-refiner loop that validates and repairs itineraries before export
• Expenses Tracker — full-stack MERN application with JWT auth, deployed to AWS EC2

Skills

• Geospatial & Data: STAC, COG, TiTiler, eoAPI, PostgreSQL/PostGIS, TimescaleDB, Xarray, Dask, Zarr, Parquet, NetCDF
• Languages: Python, C++, Java, Kotlin, JavaScript/TypeScript, SQL
• Backend & Web/Mobile: FastAPI, Node.js, Express, React, Leaflet, Android SDK, REST APIs
• Distributed & Systems: gRPC, Protobuf, concurrency (pthreads, mutexes), Docker, Unix/Linux
• Cloud & Tooling: AWS (S3, Lambda, RDS, EC2, CloudWatch), GCP (Vertex AI, Cloud Run, GCS), GitHub Actions CI/CD, Git, Claude Code
• Machine Learning: PyTorch, TensorFlow, scikit-learn, production ML inference

Education

• M.S. in Computer Science, Georgia Institute of Technology — expected Dec 2026
  • Coursework: Graduate Algorithms (Data Structures & Algorithms), Operating Systems, Software Development Process, Machine Learning, Data & Visual Analytics
• Ph.D. in Environmental Sciences, University of Virginia — 2025
• B.S. in Atmospheric Sciences, Lanzhou University — 2019

Publications

Three first-author papers on atmospheric dynamics and climate model evaluation in Geophysical Research Letters and JGR: Atmospheres. Full list, research summaries, teaching, and conference presentations are in the academic archive.