Lukas Rabe

As part of the global CRM and data governance team, I drive scalable integration between operational systems (SAP, Odoo, ForgeRock) and our Salesforce platform. I specialize in automating complex data processes, improving customer and product data quality, and enabling structured insights for segmentation, onboarding and operational alignment.

Key Responsibilities & Achievements:

Designed and implemented a custom Python-based matching engine to deduplicate and reconcile 40,000+ customer records from ForgeRock and Odoo into Salesforce — leveraging geolocation, fuzzy logic, and semantic scoring

Developed a multi-layered mapping strategy to integrate M&A customer data with product usage analysis, including automated mass onboarding and CRM enrichment

Built an automated product quality assignment pipeline using historical sales data, SKU classification, and CRM metadata — increasing segmentation accuracy across business units

Created a governance and data repair framework for SAP customer and material master data (Customer Group 4 / Material Group 4), driving alignment between direct/indirect sales structures

Engineered a Sales Office logic system for assigning regional ownership based on NUTS-level segmentation, business structure, and Route-to-Market logic — designed for automation in future onboarding flows

Collaborated closely with stakeholders across Sales Support, IT, and Regional Management to align business rules and data logic

Experienced in using Power BI for reporting and validation of customer segmentation and product usage trends

Tech & Tools:

Salesforce, SAP R3, Python, Supabase, Power BI, Excel/VBA, SQL, GPT APIs, SQLite, FAISS

As part of the Dataniva platform, I developed a modular backend for managing and deploying GPT-based AI assistants across business use cases.

Each bot is registered in Supabase with a custom config: pricing model, token limits, use case, and runtime rules. Prompts are built dynamically from database-driven templates and adjusted in real time based on user, role, and session context.

A key feature: bots can return structured JSON outputs — enabling downstream automation, such as onboarding flows, semantic product mapping, or system configuration.

All bot interactions are tracked via a custom-built LOCs (Log-of-Call Sets) framework — logging token usage, latency, context, and cost per call. This architecture allows full traceability, auditing, and per-customer billing.

Core features:

Modular GPT bot registry (Supabase)

Dynamic prompt assembly with role/history logic

Structured GPT output (JSON-ready)

Real-time usage + cost tracking

LOCs-based execution logging & replay

Support for scalable AI-driven backend processes

Impact:

This engine enables Dataniva to run auditable, cost-aware AI bots across workflows — powering automation, guidance, and product intelligence at scale.

Skills: Python · OpenAI API · Supabase · Prompt Flow Engine · JSON Rendering · Token/Cost Tracking · AI Integration Architecture

As part of the OEM-focused customer service team, I was fully responsible for managing the end-to-end order entry and fulfillment process for Porsche, VW, and selected APC clients. This included manual handling of over 300 item lines per week, often split across 3–4 different delivery addresses, each with unique production references and strict delivery protocols.

Key responsibilities & achievements:

Solely accountable for the “Color of choice” system at Porsche (“Mischbank”), enabling custom paint production from a catalog of 180+ colors, tailored to individual customers and distributed across multiple clients per batch – involving complex planning, batching, and logistics coordination.

Designed automated Excel-based planning tools using VBA and GUI scripting to streamline order creation and scheduling for Porsche’s color batches.

Collaborated cross-functionally with internal stakeholders: MPS planners, PE managers, sales teams, service managers at Porsche, plant technicians, logistics, warehousing, and IT.

Played a key role in site rollout projects (e.g., new production or warehouse sites for Porsche), including data and process readiness.

Maintained a highly fragmented product structure (180 color variants × 8 packaging sizes × multiple production sites), managing master data integrity and distribution planning.

Built and maintained planning logic to distribute paint batches effectively across 3 production sites, forecasting customer needs across different vehicle models – despite often missing or late upstream data.

Ensured compliance and readiness for audits, coordinated EDI clarification topics with IT, and continuously automated operational processes (e.g., transforming warehouse lists into shipment instructions).