Software Engineer — Backend Systems & AI Automation (via HCL Technologies) - Meta - Bengaluru, India
(2023-10)
Backend Engineering & AI Automation | Python, Distributed Systems, GraphQL, MySQL, Concurrency, AI
- Contributed to a 35-state job lifecycle state machine with granular failure classification (SSH, RPC, heartbeat, host health, timeout, framework, critical error, no-host scenarios), enabling targeted retries and improved operational debugging.
- Applied pessimistic locking via transactional SELECT ... FOR UPDATE under REPEATABLE READ isolation to coordinate multi-instance job scheduling and prevent race conditions without introducing external distributed lock services.
- Built and maintained a 4-layer retry strategy spanning database operations (5 retries), Thrift RPC communication calls with configurable retry policies, device operations with exponential backoff, and workflow-level requeue with circuit breaker guards.
- Contributed to a 2 stage queue architecture separating scheduling (10 second polling) from execution (10 concurrent worker threads), preventing slow remote I/O from blocking the scheduling loop.
- Developed backend automation tools using Python (Paramiko) for secure remote execution, improving system reliability and operational efficiency.
- Reduced test scheduling latency significantly approx (from 5 minutes to near real-time) by implementing batch insert operations, eliminating network round-trip overhead during high-volume run creation.
