Built a backend server for telegie.com using the Java Spring Boot framework, hosted on DigitalOcean. Dockerized the instance for rapid iterations. Integrated with Stripe for payment/subscription and Identity Services by Google and Apple using Keycloak as a backend.

Built an iPhone app for Telegie, utilizing both Swift and C++. Integrated with Identity Services by Google and Apple using Keycloak as a backend.

Built parallel task servers for 3D video processing on Hetzner instances utilizing Python, CUDA, and other GPU techniques.

Built and integrated a CI/CD pipeline for pypi deployment for librgbd, an open source library built by Telegie to target various versions of macOS and Linux.

Automated importing of external data to Snowflake using Airflow, Python, and defined data schema for 8 new products such as to support conversion and attribution analysis. Segment event clustering and analysis over 12M+ sessions using data mining and stats.

Migrated all 2,300+ listings by leading the design and implementation of the integration logic between Zeus and Airbnb's listing and bookings management systems using Airbnb's API + Internal tools to manage accounts/listings/bookings with Ruby on Rails & React.

Active listing count grew from 110 to 1560, peaking at 2327 in Mar 2020. (3100+ listings onboarded)

Imported 1,000+ partner listings to Zeus by designing and implementing microservices to support integration between Zeus and Rentals United using Rentals United's API and microservices on AWS, Amazon SQS, Redis, and Kafka streams.

Designed and developed an initial version of Dynamic Pricing Engine (DPE), based on seasonality and availability using Ruby and SQL.

Developed logic to compute gaps between reservations and potential loss of revenue using Ruby and Python.

Integrated Jumio and Berbix's ID verification flows and data schema using React and Ruby on Rails.

As a Data Ambassador, helped build various business metrics, such as quarterly and yearly goals, and business intelligence dashboards across all departments, using SQL, Looker, Snowflake, Python, and Ruby.

Concurrent residents staying with Zeus grew from 248 to 2700. and Monthly Billed Revenue grew from $500K to $6.3M.

Grew from operating in 10 cities/2 metros (SF + South Bay) to 147 cities/12 metros.

Helped transitioning a raw SQL repository to dbt + Optimized 100+ SQL to reduce running time from several minutes to seconds.

Developed a segmentation algorithm for detecting paragraphs, images, and text columns from scanned documents.

Implemented a system that takes photo file parameters and other information to estimate the size and aspect ratio of the scanned document.

Designed and implemented a desktop tool to help debug and evaluate the vision algorithms described above.

Designed and implemented an API server that serves as a backend for processing scanned images using RESTful and/or TCP sockets.

Built a prototype Android app that takes photos using a camera, extracts documents, fixes the aspect ratio and perspective, sends them to the API server for OCR, and displays the results retrieved from the server.

Implemented a base library that allows real-time screen capturing of the Windows desktop environment.

Designed and prototyped a network-efficient codec for remote desktop environments using GPGPU with DirectCompute/OpenCL.