Albert James Email and Phone Number

Developed a communication API that enables apps to communicate with asistant through the use of GRPC. This enabled apps to have multiturn communication with assistant, sending both UI and voice/text responses even when the app was in the background.Integrated 1p Apps with the communication framework. Created dialog trees using C++ and dart in assistant to facilitate this integration.Also built client-side infrastructure in 1p apps to integrate with the communication API.Created a design to enable the communication API to work entirely on device. This involved working with on device assistant and modifyingdialog trees to make use of on device information corpora, on device NLP and on device response generation.Rewrote the communication API in kotlin to facilitate its migration to jetpack.Developed test apps that integrated with the communication API to test multi-turn dialog along with UI and text/voice response generation.These apps were used throughout the prototyping phase of the API as well as to test new features

Redesigned payment system, reducing the development time for integrating new payment providers from a couple of months to a few weeks.Optimized payment flow (service code), stored procedures, and schema, enabling the system to handle 4800 payments per minute (over200% increase from the previous year).Integrated new payment providers and updated existing ones to handle scale, enabling over 100% growth in total payments YOY.Designed and implemented a deposit limit system, meeting new jurisdictional requirements and offering a simpler alternative to the existingdeposit limit system. Reducing the deposit limit system-based issues by 80%.Designed and implemented the option to perform batch withdrawals based on geolocation. Enabling fraud team to increase their team sizeby 3 times

Interned at Intelligent Transportation Systems Lab Worked on improving the performance of the traffic simulation systeam and analying the output of the system.Achieved speedup of 40% through parallelization of the traffic-simulation system(DynaMIT) by dividing the region into bands and using graph coloring to determine the bands to be simulated in parallel.Compared performance of simulation system for various configurations and wrote python scripts to automate the evaluation and comparison of simulation results.

Created API to store and retrieve user profiles and transaction information of over 30 million usersusing MongoDB as the database and Redis as the cache.Reduced the page load time by revising the database schema thus imoriving improve user experience. Wrote C​ # ​ based migration scripts to change existing data into the new storage schema. Developed a porting application in C​ # and migrated ​ data from MSSQL server to MongoDB and Redis without loss of data or service interruption. Identified the root cause of system crashes and poor performance under heavy traffic and fixed it using new wrapper in C# for RabbitMQ. The revised system was capable of transferring 60,000 messages in parallel.Technologies: ​ C#, MongoDB, Redis, MSSQL, RabbitMQ, GIT, SCRUM

Automated over 100 test cases for testing the middleware of the cloud-based video conferencingsoftware system. Wrote test automation scripts in JAVA using TestNG and Gradle, which drastically reduced the time spent in testing the latest features during the 3rd and 4th quarterly release of companyTechnologies: JAVA, Cassandra, MongoDB, GIT, TestNG, Gradle, SCRUM

Developed an application in ​ C++ ​ and Matlab to classify the surface leather defects. These defectswere detected by K-means clustering using images of the leather and further sub-classified leatherinto different grades.Technologies: ​ C++, Matlab