Build, run and optimize production software.

Good reasons to call.

1. The roadmap needs senior engineers faster than you can hire them.
2. A new product has to ship on foundations that won't need a rewrite next year.
3. The cloud bill grows faster than revenue, and the Datadog bill is its own budget line.
4. The pager goes off often enough that the team stopped trusting it.
5. You're on Heroku, or a platform you've outgrown, and the migration keeps slipping.
6. Agents ship more code than ever, and nobody owns how it behaves in production.

Underneath, the same few causes: early decisions that no longer fit, no visibility into how the system behaves under load, no path from here to the target.

I work through them in order of consequence, not complexity.

Focus areas

• Product Engineering

  Full-stack product work in Python and TypeScript on PostgreSQL: OAuth2/OIDC SSO, leading an API for 15M daily users, React dashboards, client SaaS builds.

• Backend & Data

  APIs and services on PostgreSQL, schema design, queues and pipelines (RabbitMQ at YouGov, replacing multi-day batch loads).

• Reliability & Scale

  SLOs, observability, incident response, capacity under real traffic. Datadog and OpenTelemetry.

• Cost & Migrations

  57% AWS cost reduction on recent client work. Datadog cost tuning. Heroku-to-AWS migrations.

Who does the work

Founded and led by Darwin Monroy, who does the engineering on every engagement. Software engineer since 2005, SRE and platform engineer since 2020; founded the SRE function at two companies.

• ~99.5% → ~99.99%

  Core-service availability as SRE tech lead at DAT Freight & Analytics.

• 40,000+

  Serverless Postgres instances on the Bit.io platform, as SRE (Bit.io was acquired by Databricks). CI cut from 40 minutes to 4.

• 15M

  Daily users on YouGov's campaign-tracking API. I led its development and production operations for years, at 99.999% through ~100x launch surges. Also built the OAuth2/OIDC SSO.

• 57%

  AWS cost reduction on recent client work. Also Datadog rollouts and cost tuning, and Heroku-to-AWS migrations.

Through a past client, my work has served CNN, Google and Meta.

When a project needs more hands, vetted senior engineers join under contract to Ideatives Inc. You approve them before they touch your systems.

• LinkedIn
• GitHub
• postmortem.io (built and run by me, 600+ public postmortems)
• ai-native-pg write-up

Fixed fee by default, scoped after we determine fit together. Project, retainer or time and materials when the work calls for it.

• Start here · from $4,500

  Cloud Cost Review

  AWS and Datadog spend: where the money goes, what to cut first, what it saves. About a week, depending on the system. Written findings and a savings plan.

• Start here · from $9,500

  Technical Review

  Architecture, reliability, observability and cost. Two to four weeks, depending on the system. Written report: risks ranked by consequence, and what to fix first.

• Start here · new builds

  Build Plan

  Before you build: architecture and stack with trade-offs, data model and auth, deploy path with a skeleton repo in your cloud, observability plan, risk register, milestones and a fixed build quote. One to two weeks.

• Targeted Engagement

  A defined problem, a fixed scope: a migration, a performance issue, a critical integration, or a system under pressure.

• Embedded Capacity

  Senior product and platform engineering inside your team, or under your brand on your agency's client work. Monthly retainer; on retainer I can carry the pager, by agreement. US invoicing through Ideatives Inc., ET to PT hours.

• New Product Build

  Architecture and stack decisions through to production launch, in Python, TypeScript and PostgreSQL. Recent client SaaS builds in Python, Go and TypeScript. Engineers join as the scope requires.

Straight answers

Why not an AI coding agent?

Agents write code. Someone still decides what's safe to change and owns it when it pages. I use the tools too: ai-native-pg is a Postgres extension for local inference.

Who will I work with?

Darwin, directly, on every engagement. Vetted engineers join per project, and you approve their access.

Tell me what you're building or fixing, and by when.