Resume
Backend, platform and SRE engineer. Plano, TX; open to relocating anywhere in the US. US work authorization, no sponsorship needed. Available immediately.
Summary
Backend and platform engineer moving toward SRE, CS graduate from UT Dallas (August 2025). I build services in Python and Go and run them myself: a monitoring SaaS in production on a VPS, and a seven-host vSphere cluster at home with three Kubernetes environments, an observability stack, an internal CA and a fleet watchdog I wrote. Comfortable owning a system from the request path down to the hypervisor it lands on, and writing up the incidents when it breaks. Everything below is project work, not employment; each figure is sourced on the linked project page.
Engineering Projects
Reliability and Platform Work on a Self-Operated Fleet
Solo. Kubernetes, VictoriaMetrics, Grafana, step-ca, Forgejo, PostgreSQL, Python
Sep 2026 to present
- Built and operate three RKE2 Kubernetes clusters (15 nodes) on vSphere with kube-vip control-plane VIPs, provisioned from the vCenter API with cloud-init; proved VIP failover by forcing a leadership transfer (about 3 s) after the first test gave a false pass. (details)
- Wrote a fleet watchdog (Python, system cron, separate host) after a scheduled job failed 970 consecutive times with one alert; now 115 signals with stable IDs, dedupe, recovery messages and a dead man's switch, each red-run before being trusted. (details)
- Stood up VictoriaMetrics, vmagent and Grafana over 77 scrape targets (23.6M rows/hr); verified all 30 dashboard panels by executing their queries, exposed etcd fsync latency with a rolling control-plane restart, and proved the remote-write queue by stopping the database for 100 s with zero sample loss. (details)
- Ran an internal step-ca CA issuing 24-hour certificates to four services with password-free automated renewal; found and fixed a renewer that had failed silently for 17 h (6h52m from expiry) and a restart hook that bounced a service 67 times a day. (details)
- Built a PostgreSQL task queue (SKIP LOCKED claims, leases with heartbeats, approval gates, capped concurrency) for long-running automated work; found a reaper bug where a NULL lease made stuck tasks invisible and fixed it with tests in both directions. (details)
- Set up nightly restic backups to separate hardware for five databases that previously had none, with a scripted restore that read back 9,849 rows; split a flapping alert that was red in 369 of 816 runs into an incident signal and a human nudge. (details)
- Deployed a self-hosted forge with server-side branch protection and six parallel CI slots on Kubernetes, validated green, red and green again before trusting it; exposed a production cluster through a Cloudflare tunnel with zero inbound ports. (details)
CheckPulse - Monitoring SaaS
Solo build, deployed at checkpulse.dev
Mar 2026 to present
- Built and deployed a multi-tenant uptime, DNS, and SSL monitoring service in FastAPI, PostgreSQL, Redis, and Celery; seven containers on a single 1.9 GB VPS behind a Cloudflare tunnel with no inbound ports exposed.
- Designed multi-region check consensus requiring two of three regions to agree before declaring an outage, trading small detection latency for a large reduction in false alerts.
- Detected and shut down abuse of the signup endpoint as a third-party email-validation oracle (~1,191 bot accounts over three months); added a CAPTCHA gate on all mail-triggering endpoints and split transactional from outbound sending domains to protect sender reputation.
- Integrated Stripe billing across three subscription tiers.
Distributed Log Analyzer
Go, gRPC
Jan 2025
- Built a master-worker system in Go that parses a 3.3 GB, 10M-line access log by distributing byte ranges to workers over gRPC and merging their results.
- Cut runtime from 36 seconds on a single worker to 6 seconds across 8-11 workers, and used profiling to identify the point where I/O and coordination overhead end the near-linear scaling.
API Gateway
Go, Redis
Feb 2026
- Wrote an API gateway from scratch in Go with an explicit middleware chain: request IDs, structured logging, JWT authentication with role-based access control, and round-robin load balancing across backends.
- Implemented two-layer rate limiting, in-memory for the common path and Redis-backed for limits shared across replicas, and exposed Prometheus metrics for request rate and latency.
Homelab Infrastructure
VMware vSphere 8, Linux
Nov 2024 to present
- Administer a seven-host ESXi cluster under vCenter 8 (132 cores, 607 GB RAM, 40 powered-on VMs as of October 2026), operated through the vCenter API rather than the web client.
- Applied least privilege to automation: a dedicated service account on a custom role whose mutating permissions are scoped to a single VM folder.
- Repaired a two-node PowerDNS pair behind a keepalived VIP: found silent replication drift and two health checks that could not fail, then proved failover (VIP released in about 4 s, 20 of 20 probes answered) and root-caused a netplan permissions outage I caused myself.
Work Experience
Employment while completing my degree, and current. Not an engineering role; listed so the timeline is complete.
Personal Shopper
Walmart Inc
July 2023 to present
- Delivered outstanding customer service by efficiently processing over 1,000 online grocery orders, achieving a 99% accuracy rate in item selection and packaging.
- Enhanced operational efficiency through meticulous scanning and organization of grocery items, ensuring prompt delivery and pickup services.
- Collaborated effectively with a team of personal shoppers and store associates, consistently meeting or surpassing daily productivity goals.
Skills
Infrastructure
LinuxKubernetes (RKE2)DockerTerraformVMware vSphereCloudflareAWSstep-ca / TLS
Backend
PythonGoPostgreSQLRedisFastAPIgRPCKafka
Reliability
VictoriaMetrics / PrometheusGrafanaAlerting designIncident write-upsBackup and restore testingCI/CD
Networking
DNS (PowerDNS)keepalived / VRRPCloudflare TunnelVLANskube-vip
Education
Bachelor of Science in Computer Science
University of Texas at Dallas
Graduated Aug 2025