19.10.2026 aktualisiert


100 % verfügbar
Senior Data Scientist
Bangkok, Thailand
Weltweit
Master in Business Intelligence / Bachelor in StatisticsSkills
Apache AirflowData AnalysisGoogle BigQueryETLPythonMicrosoft Power BISQLDatenanalytik
Senior Data Analyst with 10+ years across fintech, gaming and dating. At Down, I ran the data side of ~90 A/B tests, built trust amp; safety models (bad actors, underage users) that cut manual review from 6.5% to 4.5% of new accounts, and moved the company to dbt with daily data-quality alerts. Before that, automated reporting that saved 20+ hours/week and built attribution and customer-value models. Based in Bangkok (DTV visa), working remotely, Italian/EU citizen; tax resident in Czechia and invoicing from there.
Erfahrung
Berufserfahrungmehr als 10 Jahre
Freelancer seit2026
Sprachen
EnglishverhandlungssicherSpanishMuttersprache
Projekthistorie
- Led deep-dive analyses behind product and growth decisions: 53% of new VIP buyers end their first week without a two way conversation; a new woman comes back the next day after a first conversation, not a first match (68.4% vs 28.9%); matches go silent from friction, not the people (56.2%). The inbox changes I recommended shipped.
- Ran a photo-attribute study on 77k men (Gemini-labelled photos): effort shots convert best (gym 11.6% vs 8.0% average). Findings set the ruleset for the Smart Photos experiments.
- Built two trust and safety models: a bad-actor detector (scores accounts on their first 5 minutes; auto-bans right ≥85%) and an underage-user detector (83% of flags correct). Together they cut manual review from 6.5% to 4.5% of new accounts.
- Led the move to dbt and retired the legacy Airflow jobs. Added a daily warehouse-vs-production reconciliation alert that catches tracking losses within a day, and cut the experiment dashboard’s cost 98.6% ($19.64 → $0.27 per load). Coordinated 2 part-time analysts and onboarded new hires.
- Ran the data side of ~90 A/B tests, from launch checks to ship-or-stop recommendations. Built the experiment tooling and ran the review meetings.
Automated weekly install forecasts and return-on-ad-spend (ROAS) reports using Python and BigQuery, combining Adobe Analytics, data.ai and internal databases; delivered in Power BI, saving 20+ hours/week.
Full-time until company bankruptcy in Apr 2023; continued as freelance consultant through Feb 2024.
- Built and maintained automated data pipelines using GCP and Airflow, combining the company’s warehouse data with public geographic data.
- Designed interactive dashboards and ran analyses to support the team’s decision-making.
- Reduced client presentation prep time by over 80% through dashboard standardization and automated reporting.
- Deployed an ML model on GCP predicting the number of cars needing an airbag-recall replacement per zip code — from population density, state, and how much of the area was already covered — for operational planning.
- Supervised a part-time junior analyst for 7 months, providing mentoring and quality assurance.