Twilio CustomerAI

Twilio Engage Predictions

Twilio Engage Predictions

• Architect and lead develop in-house Machine Learning infrastructure. - Create a distributed cloud data processing framework using Go, DynamoDB, and AWS compute resources. - Build a C++ geometric data processing framework. - Build a web-based data processing/monitoring system using JavaScript with React and WebSocket. - Build a web-based data annotation system using JavaScript with React. - Create CLI for ML infrastructure using Go. - Write CloudFormation template for AWS resource deployment. - Create CI/CD workflow using QuickBuild.• Machine Learning Applications - Develop LCD display OCR model using Detectron2/PyTorch. - Develop a real-time Beat-The-AI game for a dental show with over 1,000 dental professionals, using Go, React, and WebRTC.• Medical Devices - Collaborate with medical device vendors for preliminary development and technical spec communication, including 3D printer and intraoral scanner.• R&D - Collaborated with top academic ML research center to solve challenging problem in dental industry using pix2pix in PyTorch. - Prepare dataset by aligning geometric dental models. - Implement depth map reconstruction algorithm for GAN.• Production - Create dental factory monitoring service using Go, TypeScript with React, and PostgreSQL. - Develop a image-based querying system for dental restorations using Scala, TypeScript, DynamoDB, and React.

• Created Python-based regression test and CI/CD framework.• Implemented Laplacian deformation algorithm using C++ with Eigen math library.• Implemented crease and curvature computing algorithms for half-edge mesh representation in CAD software using C++.• Developed Subsurface Scattering shader using CUDA.• Set the foundation for an in-house ray-tracing renderer using Nvidia OptiX.• Created real-time CUDA geometry shader for point cloud rendering.• Ported couple dozens Visual Studio projects to Linux using CMake.• Revamped and maintained a legacy 10-million-line dental CAD codebase.• Refactored and maintained Python embedded interpreter in CAD software.• Developed spatial acceleration structure using C++.

Ported C/C++ written Unity runtime to LG Smart TV with a team.Tuned game performance for specific platforms, including QNX and LG Smart TV.Developed new input system using C# to port Unity game from tablet to LG Smart TV.

Taught 3D Digital Graphics Scripting course with 20+ design students.Designed the course from scratch, including lecture materials, laboratory projects, quizzes, and final exam.Covered MEL and basic Python scripting for Maya.

Created tools using Linux shell scripting to efficiently maintain 60+ workstations for production.Wrote MEL rigging tools and designed graphical user interface for animators.Supported production using proprietary applications such as Maya, RenderMan, and Shake.

Optimized open-source MIDI software synthesizer written in C for target embedded system. Ported drivers, file systems, and Qt framework onto ARM architecture. Developed applications and designed graphical user interface in C/C++ for embedded system.Projects