Opty, LLC
Cloud infrastructure, security and cost engineering
Opty is a small business of senior site reliability engineers. We run, secure and cut the cost of production systems on AWS and GCP. 100% US-owned and based in Santa Clara County, serving government agencies and prime contractors in California and the Bay Area.
Company overview
Opty is a team of senior site reliability engineers who have spent years on call for production systems. We have operated a cloud estate of more than 1,000 services, built the VPN and secrets systems that companies run on, and done the engineering work behind HIPAA and SOC 2.
Our habit is to get a lot out of a small budget. We know where AWS and GCP bills leak, and we build security monitoring ourselves where a commercial licence would cost more than the system it protects. We work as a prime on small engagements and as a subcontractor to primes that need cloud, security or data engineering depth.
Core competencies
AWS and GCP at scale
Operated an estate of 1,000+ Cloud Run and Cloud Functions services. Shared VPC and IAM automation on GCP, VPC design on AWS, SFTP servers, Kubernetes. Also Azure, bare metal (OVH, Hetzner) and PaaS.
Cost optimization
VPC endpoints in place of NAT gateway traffic, local NVMe servers with redundancy for IO-heavy jobs, BigQuery physical storage billing, and long-term backups dumped to Cloud Storage.
Terraform with change control
All infrastructure in Terraform, including custom providers and local-exec wrappers. Terraform CI/CD with an approval step before apply. Server configuration with pyinfra.
Detection without the EDR bill
In-house alternative to commercial EDR/XDR: alerting on SSH access, bandwidth, failed checks in logs and syscall monitoring.
Access, secrets and VPN
SSO for internal services. Secrets management across CI/CD, production deployment and local development. WireGuard VPN and provisioning system built for a company of about 100 employees.
HIPAA, SOC 2 and NIST 800
Took part in and implemented HIPAA and SOC 2 programs. Automated checks against the NIST 800 framework. Security training and roll-out of security practices for engineering teams.
PostgreSQL at scale
Years of operating and tuning petabyte-scale PostgreSQL on bare metal: query performance, replication, backup and recovery.
Observability
Metrics, logs and alerting on the VictoriaMetrics stack, which costs far less to run than the usual ELK or EFK set-up.
Data and AI
Data pipelines that ingest and index energy sector and blockchain data. Inference engines and locally hosted LLMs, so data stays in your environment.
Blockchain operations and audits
Validator nodes for multiple networks, including Canton: onboarding, upgrades, monitoring and backups. Smart contract audits with published reports.
Differentiators
- We have operated what we build. Including an estate of 1,000+ Cloud Run and Cloud Functions services, and the on-call that goes with it.
- Low run cost by design. VPC endpoints over NAT gateways, local NVMe for IO-heavy jobs, BigQuery physical storage billing, VictoriaMetrics over ELK.
- Security without per-seat licences. We build detection on SSH, bandwidth, log and syscall signals in place of a commercial EDR/XDR contract.
- Every change is reviewed. Infrastructure lives in Terraform and is applied through CI/CD with an approval step.
- Compliance as code. HIPAA and SOC 2 experience, with automated checks against the NIST 800 framework.
- Senior engineers only. The people who scope the work are the people who do it and who answer the page.
- 100% US-owned, based in Santa Clara County. All work is done by our team in the United States, with a focus on California and the Bay Area.
- Fast response, fixed pricing. Incidents picked up within 5 to 30 minutes at any hour. Published monthly rates.
Past performance
Listia
Commercial, online marketplace
Multi-year production operations: release pipeline and time to production, monitoring, VPN and site-to-site networking, and adoption of new technology.
Monday Capital
Commercial, digital assets
Around-the-clock operation of blockchain validator nodes, including monitoring, upgrades and incident response.
References and further commercial past performance are available on request.