Senior Software Engineer at Meta rated Greatly Exceeds Expectations three consecutive years. Built a high-throughput agentic workflow engine processing 15M+ workflows per day using LangGraph, multi-agent coordination, and ReAct patterns. Delivered 500K+ mobile app downloads and 250%+ login success improvements.
Designed and built from the ground up a high-throughput agentic workflow engine processing over 15 million autonomous workflows per day, powering use cases across content integrity, policy violation detection, and model benchmarking
•
Architected the system using LangGraph as the core orchestration framework, supporting tool binding, multi-agent coordination, and ReAct (Reasoning + Acting) loop patterns for autonomous decision-making
•
Optimized system performance to maximize concurrent workflow execution on limited infrastructure through asynchronous parallel processing, batched execution, connection pooling, rate limiting, and memory optimization
•
Designed a fair weighted dequeue system to replace default FIFO queue processing, prioritizing workflows by latency requirements and business value while guaranteeing all workflows are processed without starvation, enabling efficient resource allocation across online and offline use cases
Received Greatly Exceeds Expectations performance rating for three consecutive review cycles
•
Led a cross-functional team of 20 engineers and stakeholders through an 18-month migration of the entire Facebook Business Manager user base to Meta Business Suite, modernizing a legacy 10-year-old tech stack while iterating with high-priority enterprise advertisers
•
Developed full-stack solutions for Instagram, WhatsApp, account linking, login, onboarding, infrastructure changes, and business permissions for Meta Business Suite using PHP, Hack, JavaScript, React, Python, and SQL
•
Independently conceived and executed a project that drove over 500,000 downloads to the mobile application
•
Increased login success rate from Instagram by over 250% through technical changes to the business graph architecture
•
Led a team of three engineers and increased Facebook-to-Instagram account linking by 200,000 accounts, achieving 200% of the previous value
•
Identified product gaps and revenue opportunities through SQL-based data analysis, directly informing the technical initiatives that drove the above metrics
· Hired, coached, and led 10 developers using Agile management to build a SaaS and stand-alone live streaming platform with React (later migrated to Next.JS) front end with JEST for testing
· Brought in over 1500 sign ups and 100 live viewers per event (still growing) and succeeded in helping the company get in revenue through a client contract worth $5000/month
· Led team to build a fully functional Postgres and MongoDB database on Loopback 4 (Node.js framework with GraphQL) and real-time comments, polls, and reactions with WebSocket
· Initialized and maintained backend on Kubernetes cluster containerized with Docker, and later migrated to OpenShift to better log user activity for analytics
· Developed a feature on Google Travel involving complex modeling in considering user behavior, impact on revenue, and creative logging to maximize feature
· Utilized complex web frameworks, RPCs (remote procedural call), and services using Java and JavaScript
· Completed the entire software engineering development process; wrote a comprehensive design doc, wrote unit
