Looking for a job

Aulia Ilham

@auliailham

AI Engineer at Tensium building QA standards for reinforcement learning environments that test agent soundness, difficulty, and reward-hack resistance.

Indonesia

Message

At Tensium, I built QA standards for reinforcement learning environments used to train and evaluate AI agents. I translated complex failure modes into engineering checklists covering soundness, difficulty, reward-hack resistance, and non-saturation.

At Alignerr, I contributed to RLHF workflows for agentic coding by evaluating, annotating, and improving AI-generated coding solutions. I also evaluated LLM outputs for quality at Turing and assessed factuality and helpfulness through structured rubrics at Mercor.

At DataAnnotation, I evaluated model-generated code in Linux and Bash environments, designed prompts and atomic rubrics, and reviewed deep-research outputs. My reviews addressed factual precision, logical consistency, trade-off analysis, and edge cases.

At PT. Panasonic Manufacturing Indonesia, I developed an internal ASP.NET training administration system used by 1,000+ employees. I also built data and app projects including a MagangHub scraper, a QR-based grading app, and a Java reading application.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer – RL Environment QA

Current

Tensium

Jul 2026 - Present (3 months)

Built QA standards for frontier-hard reinforcement learning environments used to train and evaluate AI agents, covering reward design, hidden scoring, oracle isolation, runtime leak probes, cheat batteries, and reproducible handoff. Turned complex agent failure modes into practical engineering checklists for verifying soundness, difficulty, reward-hack resistance, and non-saturation before rollout

Reinforcement Learning QA Reward Design Engineering Checklists

Agentic Coding – Reinforcement Learning from Human Feedback

Alignerr

Jun 2026 - Jul 2026 (1 month)

Contributed to RLHF workflows for agentic coding tasks by evaluating, annotating, and improving AI-generated coding solutions.

RLHF Evaluation Annotations Coding

AI Quality Analyst

Turing

Apr 2026 - Jun 2026 (2 months)

Performed model evaluation and quality assurance for large language model outputs.

Model Evaluation Quality Assurance LLM

Rubric-Based LLM Evaluator

Mercor

Mar 2026 - May 2026 (2 months)

Evaluated model responses as an Information Technology domain expert using structured rubrics and domain-specific criteria. Assessed factual accuracy, technical correctness, logical consistency, trade-off reasoning, and edge-case handling. Delivered detailed review notes and actionable quality feedback for real-world professional use.

LLM Evaluation Rubric Factual Accuracy Technical Accuracy Logical Consistency Trade Offs Edge Cases

LLM Factuality & Helpfulness Evaluator

Mercor

Mar 2026 - May 2026 (2 months)

Evaluated LLM outputs for factuality and helpfulness using structured rubrics and multidimensional scoring. Fact-checked outputs to identify hallucinations, logical inconsistencies, biases, and reasoning gaps. Provided annotations and actionable insights supporting model fine-tuning, alignment, and continuous improvement.

LLM Evaluation Factuality Helpfulness Fact Checking Annotations Fine Tuning Alignment

Rubric-Based LLM Evaluator

DataAnnotation

Feb 2026 - May 2026 (3 months)

Evaluated model-generated code in Linux and Bash environments across large codebases containing 200+ files. Designed prompts for complex LLM tasks and created atomic rubrics focused on factual precision, logical consistency, trade-off analysis, and unsafe or off-topic content. Audited AI-generated outputs, documented edge cases, and provided actionable feedback to improve fine-tuning data quality.

LLM Evaluation Linux BASH Code Review Prompt Design Rubric Factual Accuracy Logical Consistency Trade Offs Audit Fine Tuning Deep Research

Software Engineer / Full Stack Developer

PT. Panasonic Manufacturing Indonesia

Jan 2025 - Dec 2025 (11 months)

Developed an internal ASP.NET training administration system used by 1,000+ employees, including workflow automation, bulk Excel imports, data validation, and hierarchical approvals. Gathered requirements and translated operational needs into database schemas and automated processes.

ASP.NET Workflow Automation Excel Import Data Validation Database Design Requirements Gathering

Education

Degrees, certifications, and relevant coursework

Politeknik Elektronika Negeri Surabaya

Bachelor of Applied Informatics Technology, Informatics Technology

2022 -

Grade: 3.62

Pursuing a Bachelor of Applied Informatics Technology with a GPA of 3.62, including coursework in machine learning, algorithms, data structures, Python, and agile product-based learning.