Madrid, Spain | +34 644 477 755 | [email protected] | github.com/marcobaturan

SDET QA | AI-Driven Test Automation | Regulated Fintech

SDET QA with over 7 years of experience in regulated fintech and banking (Allfunds), delivering test automation across more than 200 bank client projects. Designed Python and Robot Framework automation frameworks integrated into Jenkins CI/CD. Applied machine learning to test generation in production from 2018, and since 2025 has built LLM and RAG pipelines for compliance and operations.

May 2026 – Present

• Verified RLHF outputs on text and office artefacts (Word, Excel, PowerPoint): style and layout errors, colour and image choices with rights or policy implications.
• Applied LLM response evaluation rubrics across accuracy, safety and coherence.

Jun 2025 – Dec 2025

• Designed and deployed a compliance system for Google Play app policy correction requests, previously handled case by case manually. LLM + RAG pipeline over policy rules, with SQL storage of cases, feedback and per-rule adjustments, designed to progressively reduce compliance team intervention as it learns.
• According to the compliance users, it freed at least a quarter of their manual review time, allowing them to focus on negotiation with Google.
• LLM and RAG pipelines on Google Cloud Platform; async Python microservices for legacy operational workflows.

May 2018 – Jun 2025

• Designed and built, together with a fellow QA engineer, a Python automation framework that delivered QA projects for more than 200 bank clients. Cycle per project went from one month to one day (two days on average), with a team of two.
• Automated generation and scheduling of content, format, functional, security (SQL and JS injection into fields), performance (JMeter) and stability tests from user stories, runnable from the command line or GUI.
• RegEx-to-Gherkin engine generating data validation scenarios for complex financial APIs; automatic verification of JSON endpoints.
• Researched and prototyped ML/AI for testing: a system that reads a front end and generates full Gherkin tests executed on Selenium, launched as a Jenkins job within the CI/CD chain. Entered production within the first six months (2018).
• Developed 25 tools automating repetitive tasks of the role.
• Applied linting, regression, integration and smoke testing across teams.

Jan 2018 – Apr 2018

• Server-side APIs for financial and public sector projects; Python chatbot for email and scheduling assistance.

Full-stack development and infrastructure support for SMB, clinic and educational platforms (Django, Python, OpenERP/Odoo) across Windows, Linux, macOS and Android.

• Languages: Python, SQL, Bash, JavaScript (production); TypeScript (studying)
• QA: Robot Framework, Selenium, pytest, Gherkin/BDD, JMeter, Jenkins (production); Locust (personal projects); pylint, bandit (studying)
• AI / LLM: LangChain, LangGraph, RAG, MCP, GenAI APIs, Groq, GCP (production); Ollama, ChromaDB, Qdrant (studying)
• Automation: Make, Zapier, n8n, AppScript (production)
• Front: React, Vite (personal projects)

• Research collaborator, University of Seville (Cryobiology Dept.): prototype for identifying lower-toxicity cryoprotectants using Python and LLM/RAG.
• Google Scholar: scholar.google.com/citations?user=gqWxbUYAAAAJ
• For education & certification, please read my GitHub profile and associated repository for this topic.