As a member of the Core team of Altruist, the Accounts team, led the Beneficiary improvements initiative, by coordinating different lines of work for ultimately reconciling data in our clearing platform and our accounts platform and also developed a daily job using Temporal and Snowflake for detecting, identifying and displaying all the different types of data points divergences between the 2 platforms and automatic code generation to fix them.

• As Principal Software Engineer for the NoteIt App, most notably re architected and implemented the whole push notifications pipeline improving accuracy in the intended delivery time as well as improving scalability and fault tolerance under Google Cloud using a mix of G-pub/sub, G-Cloud Tasks, G-Dataflow streaming jobs with change streams, G-functions, G-functions gen2, Java, Typescript and OneSignal client.

• I also migrated key data from Postgres to Spanner using Google Dataflow and migrated several services written in typescript to make use of Spanner instead of Postgres

• Successfully led the company’s top project, the multi tier org. Was responsible for the architecture design, general tech direction and most of the implementation of the backend and worked extremely close with product, the frontend team, the backend team and DevOps.The project was based on Spring Boot and Spring WebFlux, API driven using SwaggerHub as a means of contract between frontend and backend, developed under AWS and integrated with Aurora DB.

• Also led the roles shift system re architecture project which relied on a heavily decentralized architecture with profuse use of AWS Lambdas and queuing technologies to form a highly cohesive loosely coupled system.

Completed new search service based on Spring Boot and Spring WebFlux which used the latest version of Elasticsearch to perform indexing and searches. Added new features to our own erlang client for pulsar and made use of it through different parts of the system to improve scalability and fault tolerance like for example making indexing requests go through pulsar leveraging the retry mechanism and adding back pressure through pulsar throttling feature