AI consulting, platform engineering consulting & software consulting services

Hands-on help for startups: production systems, cloud infra, and LLM-backed products — not generic strategy decks.

I work as a fractional platform engineer and AI infrastructure consultant for startups that need senior help with Kubernetes, GCP, Azure, DevOps, GitOps, and production LLM pipelines — without committing to a full-time hire yet. Background: Gojek internal PaaS, sole founding engineer at Likeo (acquired), multi-cloud migrations and high-throughput LLM systems. More on my main portfolio.

Who this is for

  • Seed–Series B product companies shipping software in the cloud.
  • Teams stuck on infra, migrations, or an LLM pipeline that is fragile or too slow.
  • Founders who want an interim owner for platform or automation, not slide decks.

Software consulting services: platform engineering, DevOps & LLM work

  • Cloud & platform — multi-region or multi-cloud moves, Kubernetes / GitOps, CI/CD, observability, cost and reliability hardening.
  • LLM & automation — production pipelines (evals, throughput, guardrails), workflow tools, integrations with your stack.
  • Internal products — developer platforms, admin tooling, data/event pipelines your team relies on daily.

How engagements work

  1. Intro call — scope, constraints, success criteria (book a free intro call on Cal.com).
  2. Short proposal — milestones, time window, and commercial terms.
  3. Execution — async updates and weekly sync; you keep ownership of repos and accounts.

Typical windows are a few weeks to a few months, depending on depth. I am also open to selective founding engineer conversations if the fit is mutual — this page is for explicit consulting and fractional work.

Related writing

Longer context on how I think about platform work, AI in production, and infra—before you book an intro.

  • AI InfraLightweight AI pipelines on schedules—GitHub Actions as the runtime instead of extra servers.
  • Multi-cloud API keysCentralizing third-party secrets across dozens of services without rewiring every repository.
  • Steer the LLMProduct and org intent guiding LLM behavior—when humans steer the model vs. the reverse.
  • Learning Kubernetes the hard wayHands-on cluster notes, constraints, and gotchas from working through real setups.

Get in touch

Book a free intro call, or email a short note (problem, timeline, budget band if you have one). I reply within a few business days.