İstanbul, Turkey
Ersan Sevinç
Technology Consultant & Solution Architect · Insurance, Automotive, ERP & Regulated Industries · AI-Native Delivery
I help organisations in insurance, automotive, ERP and other regulated industries turn complex systems into working products — from requirements and solution architecture to AI-native delivery. I currently lead the delivery of a new insurance product across 12+ integrated systems, and own an AI-native training platform that went from mockup to 450+ users in three months.
What I deliver
Architecture to production
I shape the solution, model it, write the requirements and see it shipped. My current platform went from a working mockup to 450+ users in three months, retiring the legacy system it replaced.
AI that holds up in production
LLM-assisted delivery with validation built in, and RAG / GraphRAG platforms that run on the client's own models when data cannot leave the building.
Delivery in regulated domains
Insurance, ERP and compliance work where traceability, audits and UAT are part of the job, not an afterthought. PMP® and ITIL® certified, with ISO 27001, TISAX and ISO/IEC 33002 (SPICE) process assessment experience.
Selected work
AI-Native Training Management Platform
2026 – presentEnterprise product · 450+ users
Originated the concept, secured executive sponsorship, and took it from mockup to production in three months. Now the system of record for training across the organisation.
Next.jsReactTypeScriptTailwind CSS.NETClean ArchitectureGraphRAG Knowledge Platform over Legacy Mainframe
2026Generative AI · Global insurer · 70+ users
A generative-AI knowledge layer over decades-old PL/1 mainframe systems, turning undocumented tribal knowledge into something any developer or analyst could query.
RAG / GraphRAGLLM & embedding modelsPL/1JCLBatch schedulingGraph modellingEnd-to-End Insurance Product Delivery
2025 – presentInsurance technology · 12+ integrated systems
A new return-of-premium insurance product, delivered turnkey for a global insurer across its core and peripheral systems — from actuarial and underwriting to payments and dunning.
ScrumRequirements engineeringGap & impact analysisAPI testingUATPostman
How I work
- In the code
- I own the roadmap and stay in the repository. Specs get tested against what actually ships, not handed across a wall.
- Testable specs
- Requirements are written so they can be traced to a test and an outcome. That is what lets a regulated team ship with confidence.
- Validation first
- With AI, generation is the easy part. I treat hallucination risk as an engineering concern: guidelines each sprint, review, and test gates.
- Evidence over opinion
- An audit background means I measure adoption, find the root cause, and show the result, whether that is a KPI review or a go/no-go call.
Get in touch
Working on something in AI-native delivery, insurance technology or enterprise modernisation? Send a message.