AI is only as good as the judgment behind its data.

I evaluate and rank AI responses, audit annotated data, and check Indonesian language output the way a native speaker actually reads it. Freelance, remote, for AI teams that need Indonesian coverage they can trust.

Luthfi Septian, Jakarta, Indonesia. Native Indonesian, English B2. Available 40+ hours per week, including full-time contracts.

Send a project brief View LinkedIn profile
Portrait of Luthfi Septian
Luthfi Septian, Indonesian LLM evaluation and data annotation QA.

Services

Four kinds of work, all in the same craft: reading carefully, deciding consistently, and writing down why. Every delivery includes the reasoning behind the labels, not just the labels.

01

AI response evaluation and ranking

Side by side comparison of model answers, ranked against your rubric: correctness, instruction following, tone, and Indonesian naturalness. This was the core of my Outlier work, including answers delivered and assessed by voice.

02

Data annotation and QA review

Classification and labeling done to a written guideline, plus a second pass that audits other annotators: error patterns, borderline cases, and guideline gaps flagged in a short QA report.

03

Indonesian language work

Naturalness checks on model output, Indonesian to English translation review, and speech and transcription review. Native judgment on register, politeness levels, and phrasing that reads translated even when the grammar is correct.

04

Search and content evaluation

Rating search results and content against relevance and quality guidelines for the Indonesian market, with notes an engineer or PM can act on.

Experience

Paid platform work and current training, stated plainly.

AI Generalist, Outlier (Project Aether)

Freelance, August to October 2026

Evaluated and ranked responses from two models across text and audio tasks in Indonesian and English, working to project rubrics with voice delivered answers.

RWS TrainAI, Project Callisto (training)

Learning phase completed November 2026, Indonesian language test passed

Search quality evaluation training. Client exam enrolment is the next step, so this is listed as training, not employment.

IBM SkillsBuild, Artificial Intelligence Fundamentals

In progress, started November 2026

Part of a staged path toward a verified AI Specialist credential, followed by a specialist portfolio in Indonesian LLM evaluation.

Assistant Store Manager, Alfamart

May 2025 to June 2026

Retail operations and administration. The unglamorous training behind my consistency: checklists, audits, and numbers that have to reconcile at the end of the day.

How a project runs

1Brief

You send the task, the rubric, and a deadline. I reply with questions where the guideline is ambiguous, before any work starts.

2Pilot sample

A small paid sample first, so you can check my judgment against yours and adjust the guideline while it is still cheap to change.

3Production

The full batch at the agreed pace, with progress notes instead of silence.

4QA report

Delivery comes with error patterns and borderline cases written up, so the next batch starts smarter.

Rate

One rate, no tiers, no surprises.

US$15 / hour

Negotiable for steady volume. Pilot samples are billed at the same rate. Payment by bank transfer or your platform of record. Jakarta based, working hours overlap Europe and Asia comfortably, with agreed overlap for US teams.

Project brief

Fill this in and it opens a ready to send email in your mail app, addressed to me. Nothing is stored on this page.

Prefer direct contact? Email luthfiseptian1303@gmail.com, WhatsApp +62 815-8454-3140, or message me on LinkedIn.