Dimas Novandra Utomo

What I actually do #

I'm a Technical Account Manager at ICS Compute, where I lead the technical relationship for a portfolio of enterprise customers across Indonesia. Most of my week happens inside a triangle:

• Customers who need vendor capabilities to map cleanly onto their architecture, compliance posture, and engineering culture
• Vendors whose roadmaps move faster than enterprise procurement can absorb
• Engineering teams who have to land the integration — and own it at 3am

The unglamorous middle is where most of the value lives. I do architecture reviews, vendor evaluations, escalation triage, and the diplomacy of getting things shipped at the speed enterprise customers actually move.

How I got here #

I trained as a DevOps engineer and spent the better part of a decade in production systems — Linux, Kubernetes, observability, the plumbing that decides whether your application gets to be reliable or just exists. That background still shows up in everything I do now: every architecture review starts from "how does this fail" rather than "how does this work."

AWS is where most of my customer work lands — it's the depth I keep current. But the interesting problems are vendor-agnostic at the architecture layer: how data moves, where state lives, what fails when, what costs what at scale. The cloud provider matters less than how you reason about it.

The pivot to AI happened naturally. When a customer asks "can we run this LLM workflow ourselves?", they're asking an infrastructure question, not a model question. There's a generation of platform engineers about to discover that the hardest part of "agentic AI" isn't the agents.

What I'm thinking about #

• AI in production at enterprise scale — vendor consolidation, the demo-to-prod gap, when to buy vs build, what "agentic" actually costs to run
• Cloud as a discipline, not a brand — AWS-deep where it counts, but the lessons port across providers
• Platform engineering as leverage — developer experience as the difference between AI initiatives that ship and the ones that get rewritten every quarter
• The TAM craft — technical depth as a moat in B2B, why most "customer success" misses what enterprise customers actually need

Outside the day job #

I run a homelab from Jakarta — Proxmox on an old ThinkPad, a dozen LXC containers, OpenVPN tunneled out to the VPS this site runs on. It's how I stay in the mud while my day job moves further into strategy.

This site itself is a small example: FastAPI + Postgres + pgvector for hybrid semantic search, a Linear-style design system, deployed on a 1GB Linode alongside a decade of legacy projects. Built in an afternoon — because the muscle is still there.

Available for #

• Fractional advisory — AI infra strategy, vendor reviews, architecture sanity checks
• Technical writing collaborations on the topics above
• Speaking — meetups, internal customer days, podcast guests
• Friendly conversations about hard problems — especially the ones nobody else wants to own

Drop a line at dimasutomo@dimasutomo.com or via the contact form. I read every message and reply personally, usually within 48 hours during the workweek.