Idzharul Huda

Analytics and Reporting Lead

Rey.id - Jakarta, Indonesia | Aug 2026 - Present

• Lead the analytics and reporting function and own the batch analytics platform behind client reporting.
• Run 25 ingestion pipelines (21 hourly, 121 tables, about 2,900 table loads a day) landing into BigQuery through Cloud Storage with a change-set merge and a 2-hour look-back window.
• Maintain 599 dbt models and 908 automated data tests; 233 models rebuild hourly, 124 daily, all transformation as SQL inside the warehouse, run from git through GitHub Actions.
• Added a daily completeness check that compares what each pipeline claims it built against what BigQuery actually changed, so a job that reports success while writing nothing is caught.
• Keep the platform fully on-demand - 3.5 TB scanned a day (about 105 TB a month) with no capacity reservations - and deliver client reports as staged files over SFTP or email with per-contract password archives.

Data Analyst (Analytics Engineer)

Rey.id - Jakarta, Indonesia | Aug 2022 - Aug 2026

• Optimised a key datamart model to run 9x faster, cutting BigQuery slot time by 96% and data shuffled by 58%, contributing to a 30% gain in overall pipeline efficiency.
• Raised report accuracy from 82% to 95% by automating insurance claim workflows with dynamic master mapping.
• Migrated reporting from Google Sheets to BigQuery with dbt, moving report generation from daily to hourly and enabling 40+ stakeholders to self-serve.
• Built end-to-end data quality checks and automated discrepancy detection, cutting manual error resolution by 50%, and enforced ISO 27001:2022-aligned data handling.
• Enforced personal-data controls in the platform itself: column-level policy tags on name, email, contact number, address, ID and financial fields, 29 tagged models, and a 776-table / 1,473-relationship catalogue that blocks personal data from readers without clearance.

Data Analyst

Rize Farm - Jakarta, Indonesia | Jul 2026 - Aug 2026

• Built an hourly multi-source sync that pulls eight analytics cards into a single reporting workbook, with the SQL for every card version-controlled in the repository.
• Added a divergence layer that freezes closed weeks and records only real changes, retiring thousands of rows of already-known state that each run used to rewrite.
• Shipped a read-only Apps Script dashboard for weekly per-person and per-team incentive metrics, with a diagnostics panel that reports a missing or renamed column instead of returning silent zeros.
• Moved all aggregation into the workbook so the app only renders, and migrated a card that outgrew the query endpoint row cap onto a full export with a written runbook.

Data Analyst

Ministry of Education and Culture, Republic of Indonesia - Jakarta, Indonesia | Nov 2023 - Nov 2024

• Led statistical analysis with Winsteps Rasch modelling to improve the accuracy of student talent and interest categorisation.
• Built a multi-level difficulty assessment pipeline in Python for nationwide student fitness assessments, and standardised the evaluation norms adopted in the national feasibility study.
• Authored replicable project documentation so the method could be re-run by the next team.

AI Engineer

Olvo.ai - Hong Kong, China (Remote) | Apr 2024 - Jan 2025

• Digitised thousands of hardcopy documents into structured datasets with OCR, YOLO and GPT-4, improving document handling efficiency by 70%.
• Designed an LLM embedding system to standardise hospital formularies across institutions, and automated claims verification and pricing audits, cutting manual review by 60%.
• Built a real-time client demo with Streamlit for the client evaluation.

Bachelor of Computer Science

Sebelas Maret University - Surakarta, Indonesia | Aug 2018 - Mar 2022

Coursework in programming, software development and problem solving; thesis on automated essay scoring (deep learning with Word2Vec and TensorFlow, 91.23% kappa on the Kaggle Hewlett Foundation dataset).

Dataverse - Conversational Analytics

Public repository | github.com/idzharulhuda13/dataverse

• Multi-agent app that answers questions over uploaded CSV, Excel and Parquet datasets: cleaning, enrichment, SQL generation and visualisation specialists with 191 commits shipped.

Automated Essay Scoring (thesis)

Python, Word2Vec, TensorFlow | 2022

• Trained a deep learning scorer on roughly 1,700 essay samples and reached a 91.23% kappa score on the Kaggle Hewlett Foundation dataset.

Telehealth Cost-Effectiveness Study

Published research | 2023

• Showed telehealth-powered managed care was 7x more cost-effective than conventional care for acute respiratory infection.