Data & AI Advisory · 2026

Turn messy data into decisions — and ship AI that earns its keep.

I work with mid-market teams as a fractional Head of Data & AI. Six-week engagements, clear scope, no slide-ware.

Book a 30-min intro call → See how it works
7+ yrs at Eureka AI (Singapore) — Data Scientist → Head of Engineering & ML Creator of Seeknal · Co-founder ClaimMind (Antler ID9)
How it works

Four phases. From Discover to Embed.

Same shape every time — no 40-page kickoff deck. Three sequential phases take one use case from messy data to production; then an ongoing phase if you want me to stay on the team. We tighten scope on the intro call.

Phase 01

Discover

A 30-min intro call, then a short diagnostic. Stakeholder interviews, a system walk-through, and a hard look at the data you actually have — pipelines, warehouse, feature store, and ML infra seen through a senior engineer's eyes.

You leave with

  • Architecture diagram of current state
  • Risk & cost scorecard (reliability, latency, cost)
  • An honest read on the data & systems you really have
Start here →
Phase 02

Diagnose

Cut through the agentic-AI hype. I score your readiness, rank real use cases by feasibility and ROI, and tell you what's broken, what's working, and which two bets are worth making this quarter.

You leave with

  • Agentic readiness scorecard (data, talent, infra, governance)
  • 2 ranked use cases with feasibility & ROI sketch
  • A clear go / no-go verdict
Book a call →
Phase 03

Deliver

From whiteboard to working agent. We pick one use case, build it end-to-end on your data, and hand it over with the eval harness, guardrails, and runbook a real team can operate. Code, not slides, wherever it makes sense.

You leave with

  • 1 production-ready agent (your stack, your data)
  • Eval suite + safety guardrails
  • Operating runbook for your team
Book a call →
Ongoing After that

Embed

I don't ghost you after the kickoff. As a fractional Head of Data & AI, I stay on your team — running the cadence, coaching your AI champions, and shipping. Not a 6-month enterprise rollout, but a partnership that runs as long as you need it.

You leave with

  • Ongoing fractional data & AI leadership
  • A steady cadence: roadmap, review, and execution
  • A team that levels up, not one that leans on a vendor
Talk partnership →
About

I'm Fitra. I build the things I advise on.

I spent seven years at Eureka AI in Singapore — joined as a Data Scientist, left as Head of Engineering & ML — building production data and ML systems for telco-scale workloads. I'm the creator of Seeknal, an open-source data pipeline with a built-in AI agent and semantic layer, and co-founder of ClaimMind (Antler ID9), an AI revenue-cycle tool for Indonesian healthcare.

I write code, run workshops, and don't ghost you after the kickoff. My bias: ship something a real user can poke at by week two, then iterate. No 6-month enterprise rollouts.

  • Faculty, Bina Nusantara University (since 2011)
  • MTI, Bina Nusantara University
  • Published in bioinformatics, ML, cloud computing
  • Based in Tangerang · works globally remote
FAQ

Questions worth answering.

How do we work together?
Three sequential phases — Discover, Diagnose, Deliver — with scope and deliverables fixed and agreed on the intro call. For longer-term support there's the Embed phase: a fractional Head of Data & AI who stays on your team. We finalize scope on the 30-minute call.
How long is a typical engagement?
Discover through Deliver runs in days to weeks — one use case to production in roughly six weeks at most. Embed runs as long as you need it. If a vendor is pitching you a 6-month "discovery phase," walk away.
Do you write code, or just slides?
Both, but my bias is code. I've spent more time in pull requests than in PowerPoint. Where a working prototype beats a deck, you'll get the prototype.
What industries do you work in?
Strongest in telco, fintech, and healthcare — that's where Eureka AI and ClaimMind put me. Anywhere with messy operational data and an exec who wants real answers is a fit.
Do you sign NDAs and work with regulated data?
Yes. Mutual NDA before any sensitive walkthrough. I've worked with PII at telco scale and PHI in healthcare; least-privilege access and synthetic-data demos by default.
What if I don't know what I need yet?
That's what the 30-min intro call is for. If it's not a fit, I'll tell you on the call and point you to the right person — including "you don't need a consultant yet."
Book a call

A 30-minute call. No deck. No pitch.

Tell me the one data or AI question keeping you up at night. I'll tell you where to start, which phase fits, and what I'd do in your seat. If we're not a fit, you'll know inside thirty minutes.

Limited to 8 slots / week · 15-min buffer · Pre-call form filters tire-kickers.

Intro call · 30 min

Video · 30 minutes · pre-call form
  • FormatZoom call
  • Length30 minutes
  • OutcomeFit verdict + next step
Pick a time on Cal.com → Email instead