I'm Vaibhav — a systems engineer working across Kubernetes, CI/CD pipelines, and distributed backends, plus the full-stack applications that sit on top of them.
Accel-backed IoT device-management platform — OTA updates, fleet monitoring, and analytics for connected devices.
Eliminated a CWE-78 command injection vulnerability shared across 4 CLI commands by replacing unsafe shell string-building with direct process execution; added regression tests.
Built a health-monitoring service on a public-facing bootstrap VM that SSHes into 20 private EC2 machines every 5s to collect CPU/disk metrics for 30+ services, storing results in DuckDB with a rolling 24-hour retention window.
Contributed to the team's internal CLI (5-person team) for VM provisioning; added persistent SSH remote-attach via syscall.Exec with atomic tmux attach-or-create, eliminating duplicate-session races.
Open-source contributions to Nomad's scheduler, client, and server-side authorization.
PR #27880 ↗Fixed NUMA-aware scheduler bug causing false node exhaustion from stale device-allocator state after preemption; added regression test.
PR #28191 ↗Fixed client allocation watcher to stop retrying on permanent RPC errors (permission denied, unknown allocation) instead of looping and flooding logs; added regression test.
PR #28110 ↗Fixed server-side authorization to let a node sync its own allocations after a node-pool change, by extending AuthorizeClientAllocation to bypass pool/namespace checks on matching node ID.
Go · Nomad · Distributed Systems · RPC · Regression Testing
projects
Things I've built
March 2026
Faultlab
FultLab is a distributed fault injection system that programmatically introduces failures across nodes to evaluate system resilience under real-world conditions.
Blacktree is a container-based deployment engine that builds, schedules, and runs applications from source code with automated orchestration and execution.
Containerization · DevOps · Job Scheduling · Distributed Systems
