Café Corner
A little corner where I collect my café explorations
+
Experienced in cloud-native architecture, data engineering, and analytics within global financial enterprises. I build scalable, reliable systems that empower people to make better decisions.
Explore My WorkImplemented a Foundry AIP + GPT-5 pipeline for PDF treaty contract comparison, including PDF ingestion and normalization, batched LLM inference with prompt templates, and clause-level change detection. Exposed results via an LLM-driven analytics workflow for underwriters and delivered an internal Python integration library to enable rapid reuse across P&C pipelines.
Led the design and deployment of scalable data pipelines in Palantir Foundry and PySpark to ingest and harmonize multi-year exposure data from 450+ cedents. Delivered the Casualty Exposure Analytics Suite, integrating exposure metrics, loss model outputs, and underwriting data into a unified analytical layer supporting renewal and pricing decisions. Introduced automated risk-monitoring components to detect emerging exposure trends, enabling underwriters to identify portfolio shifts in near real time.
Production-grade Go API platform on Azure AKS using Helm, Terraform, and Docker. Integrated new data sources, improved system resilience, and aligned implementation with DevSecOps best practices. Extended a Python SDK used by multiple engineering squads to query over 11 million entities daily, enhancing scalability and integration. Provided platform insights and performance reviews to senior P&C leadership to inform data strategy.
Audited, refactored, and standardized legacy projects across the P&C organization. Authored a comprehensive engineering framework outlining best practices for unit testing, regression testing, UAT, and PySpark optimization. Defined measurable metrics and a clear roadmap for progressive refactoring, ensuring consistency, maintainability, and performance across analytical data products.
Led the engineering effort for the Analytics Data Model (ADM)—a central, certified data platform consolidating data from pricing, costing, and claims applications. Enriched business data with internal reference attributes, treaty metadata, and underwriting classifications to create a consistent source of truth. Built on Palantir Foundry and SQL Server, ADM provides a governed, self-service environment for ad-hoc analysis, operational reporting, and advanced modeling across the global reinsurance portfolio.
Sole data engineer driving the development of a forecasting and risk-monitoring suite for Swiss Re Institute’s Life & Health division. Designed and deployed the Early Warning Sign dashboard, now used for real-time portfolio risk assessment. Built PySpark pipelines and data integrations in Palantir Foundry, ingesting extensive external sources such as Merative’s MarketScan data to strengthen research analytics and predictive modeling capabilities.
Engineered ETL pipelines consolidating ERP data from global dealers and customers to support sales analytics and forecasting. Developed an Airflow-based DAG workflow to automate extraction, transformation, and loading into Snowflake, significantly improving reliability and scalability. Standardized and cleansed data to empower data scientists in building dashboards and training machine-learning models for demand forecasting and trend analysis.
Developed a lightweight Flask AppBuilder web application enabling dealers to submit structured data and provide real-time feedback on processing logic. The feedback loop accelerated validation cycles and reduced re-work in subsequent ETL stages. Acquired proficiency in Flask AppBuilder, OAuth, and full-stack deployment through self-directed learning, demonstrating adaptability and rapid technical upskilling
Designed and developed a web-based technical documentation platform for the Aftermarket Engineering team, serving as a central hub for product manuals, engineering notes, announcements, and release information. Automated content population workflows and administered AWS Linux servers, including setup and teardown of hosted services. The site, built from scratch with HTML, Bootstrap, and Nginx, provided a unified entry point that improved knowledge sharing and reduced duplication across engineering teams.
Focused on data analytics, systems design, and enterprise informatics. Master’s thesis Devised Analysis Methodology for understanding the state of Retirement Preparedness of the Swiss Population administered visualised via an interactive web-based application.
Concentrated on distributed systems, software engineering, and algorithm design.
Fundamentals of GenAI and deployment and optimisation strategies
Build RAG applications using vector databases and advanced retrieval patterns
Performance optimization, automation, and secure data sharing on Snowflake’s cloud data platform.
Fundamentals of Snowflake and scalable data pipelines
Cloud infrastructure, governance, and DevOps principles within Azure ecosystem.