Reality Labs Blame Finder Tool

• ---

• Maintained and scaled platform for finding commits that caused breakages to End to End tests across all AR / VR / Wearables devices and platforms

• Used 8 programming languages and 35 tools to build solutions and increase platform coverage, bringing the success rate from 20% to 60% while scaling traffic by 18x

• Built and managed data pipelines to automatically highlight important issues and metrics

GPT Prompt Discovery Tool - https://www.sgpt.app/

• Bootstrapped AI Text Generation application and devops, including automatic CI/CD, error monitoring, database backups, and user flow tracking with Mixpanel

• Utilized GraphQL and Postgres to serve React web application with Next.js on DigitalOcean

• Example Google boolean keyword generator for recruiters: https://www.sgpt.app/prompt/urAxsh

Goodreads.com Discovery and Community Team

• Built and maintained core features used by 10 million monthly active users spanning 11 microservices architecture and 3 technology stacks

• Routinely fixed issues spanning 11 CI/CD pipelines by upgrading packages, addressing vulnerabilities, analyzing monitoring metrics, and running load tests to ensure stability at high levels of traffic

• Performed bi-monthly on-call rotations, managing high and low severity ticketing for all services

• Consolidated and deprecated 2 microservices with low usage, reducing team’s operational overhead down to 9 microservices with minimal customer impact

I'm responsible for providing technical and emotional support to a group of ~40 students. I monitor their progress, giving them extra support if they are falling behind. In addition, I try to help build their confidence and resilience in the face of challenges, to help them become happier and more productive in their professional careers.

• Horizontally scaled back-end among AWS T2-Micro instances

• Stored data in PostgreSQL due to its optimizations on complex SQL queries, maximizing performance for data retrieval from 5 tables

• Performed stress tests and analyses with k6.io and NGINX Amplify, simulating high traffic load to identify performance bottlenecks

• Increased request per second (RPS) throughput by 600% with implementation of NGINX cache and PostgreSQL hot standby

• Reached 727 RPS with an average latency below 20 ms