Results, not promises.
Every project we take on has a before and an after. Here's what the after looks like.
BlackTide - Monitoring at scale
BlackTide had built a promising Web3 monitoring platform tracking RPC endpoints, blockchain nodes, and smart contract events across 24+ networks. As they scaled to 10M+ daily checks, their infrastructure began to crack alert fatigue was constant, deployments were manual and fragile, and there was no incident response process in place.
WHAT WE DID
- CI/CD pipeline with GitHub Actions and Docker
- Prometheus + Grafana observability layer
- 6 incident runbooks documented
- PagerDuty alert routing with on-call schedules
- Kubernetes autoscaling policies
STACK
- Go, PostgreSQL, Redis
- Docker, Kubernetes
- Prometheus, Grafana
- GitHub Actions, PagerDuty
The engineering team went from firefighting to shipping and the company was able to onboard enterprise clients with confidence in their SLA commitments.
Chollofi - Performance at peak traffic
Chollofi is a fast-growing Spanish deals aggregator whose revenue depends entirely on affiliate conversions. During peak traffic windows the site was experiencing load times above 8 seconds and intermittent server errors costing real affiliate revenue on their highest-traffic days.
WHAT WE DID
- Migrated to dedicated VPS with tuned workers
- Redis object caching for DB queries
- AWS CloudFront CDN configuration
- PostgreSQL query performance tuning
- Uptime monitoring with Slack alerting
STACK
- Python, PostgreSQL
- Redis, AWS CloudFront
- Dedicated VPS
Chollofi went from a fragile site struggling under traffic peaks to a lightweight, cacheable platform loading in under 1.5 seconds with a direct uplift in affiliate conversion rates.
Cezilia - Production-grade from day one
Cezilia automates payroll processing for Spanish SMEs. When they came to us they were running on a single EC2 instance with no environment separation, no automated backups, and deployments requiring manual SSH access. They had just closed a seed round and needed infrastructure ready for enterprise clients.
WHAT WE DID
- Multi-environment AWS setup with Terraform
- Docker + AWS ECS with autoscaling
- Full CI/CD pipeline in GitHub Actions
- Encrypted PostgreSQL backups to S3
- AWS Secrets Manager for all credentials
- SOC 2 compliance readiness report
STACK
- Python, React, PostgreSQL
- Docker, AWS ECS
- Terraform, GitHub Actions
- AWS Secrets Manager, S3
The engineering team can now ship daily without risk of data loss or unplanned downtime — and the platform is positioned for a future SOC 2 certification process.
ApplyOK - Visibility into the AI pipeline
ApplyOK generates AI-powered cover letters for job seekers. Despite growing traction, the platform had no visibility into how their OpenAI integration was performing. Errors surfaced silently, users hit generation failures without feedback, and there were no rate limiting mechanisms a single power user could exhaust the token budget for everyone.
WHAT WE DID
- Structured logging with Pino end-to-end
- Sentry error tracking for OpenAI failures
- Per-user rate limiting via Supabase
- Health check endpoint and status page
- CI/CD with GitHub Actions and Vercel
- Internal token usage dashboard
STACK
- Next.js, Node.js
- OpenAI API, Supabase
- Vercel, GitHub Actions
- Pino, Sentry
The team can now detect, diagnose, and resolve issues in real time rather than discovering them through user complaints. The CI/CD pipeline allows the solo founding engineer to ship confidently at speed.
Your infrastructure could be next.
If any of these cases sound familiar, let's talk.
Book a discovery call →