Yuchuan Gou Email & Phone Number

Foster City, California, United States

San Francisco Bay Area

The main contributor to AI+ Remote Sensing project.(1) Accelerated model with TensorRT, achieved x6 throughput on Deep Learning tasks with FP16 Quantization.(2) ML Infra Development: ML workflow management with Mlflow; ML Env management based on docker. Implemented multiple-processing GPU scheduler package for highly scalable model Inference.(3) Built backend async API services with FlaskCelery queue on Pytorch Super-Resolution models to 4x scale-up satellite RGBN/DEM/Radar images. Developed pre/post processing, prediction pipeline for geo-image inference. (4) Implemented a relational database on PostgreSQL + Postgis in Spatial-Temporal Asset Catalog (STAC) standard to help fast and efficient querying global Sentinal satellite image metadata. Developed a Python client library with Peewee ORM for user's geospatial-temporal-properties query.(5) Led research work on image enhancement models for satellite images. The Super Resolution model can 4x scale-up RGBN/DEM/Radar images with SOTA level PSNR performance. The Image Translation model can predict RGB images given Radar images in cloudy weather conditions. (Winner of CVPR 2022 MultiEarth Workshop)The main contributor to AI+ Art NFT project.(1) Developed the full-cycle pipeline for AI painting generation engine to support AI + Art NFT product.(2) AI drawing board app(Siggraph2020 Poster, second author)(3) Research on GAN model quality improvement(Best Paper in CVPR2020 AI4CC Workshop, first author)

Shanghai City, China

Conducted several CNN model inference testing (from different Intel deep learning backbone) for Intel’s customers.Contributed to Intel/ Theano Github repo: 1) Wrote 2 Theano operations with Intel MKL library (C, Python). 2) wrote “Theano-perf” module for testing performance, including benchmark, CPU, disk, memory testing.

Gainesville, Florida Area

Developed an official website for Florida plants and butterflies. Realized login, bookmark, display, edit functions based on SQL database with C#. (jQuery, ASP.NET MVC, C#, SQL Server) https://ffl.ifas.ufl.edu/butterflies/Implemented a geo-locating web app to display the fertilizer map of Florida. Applied Google Maps API and built responsive web UI and map components with Materialize, JavaScript. https://ffl.ifas.ufl.edu/Fertilizer/

Shanghai City, China

Implemented Louvain community detection algorithm on Spark, performance better than Java standalone program, efficiency can be better with larger Spark cluster. (Scala, GraphX)

Master'S Degree, Computer Science

Bachelor'S Degree, Information Engineering