About
CTO | Technology Strategy • Engineering Leadership
I'm Ibrahim, a technology executive based in Istanbul. I've spent over two decades building and leading engineering teams, helping organizations scale their technology, and solving complex problems at the intersection of business and engineering.
Currently, I'm the CTO at Avansas.com, managing a diverse technology organization that spans Product, UX, Data, Enterprise Applications (CRM, SAP, etc.), Software Engineering, Infrastructure, and Security—a role that effectively blends CTO and CIO responsibilities. Before that, I held similar roles at TikTak and spent a decade at Trendyol Group, Turkey's largest e-commerce and technology company. It was there that I experienced massive scale and architectural innovation firsthand, building and managing multiple development teams focused on logistics, warehouse management systems, and back-office operations.
My work revolves around platform modernization, microservices architecture, and building systems that handle scale. I care about aligning technology strategy with business outcomes, fostering ownership in teams, and making sure engineers have the tools and autonomy to do their best work.
What I focus on
- Building and leading high-performing engineering teams
- API-first design, microservices, and Kubernetes-based infrastructure
- Warehouse management systems and logistics technology
- Digital transformation and operational efficiency
- Coaching leaders and fostering a culture of continuous improvement
- Enterprise AI Strategy and On-Premise LLM Architecture
AI Engineering
I don't just use AI tools; I build on-premise AI systems. I design LLM-powered workflows with a strong focus on data privacy, reliability, and enterprise constraints:
- Text-to-SQL Pipelines: Converting natural language questions into SQL for automated analytics with local LLMs (Ollama, Qwen2.5).
- RAG + Hybrid Retrieval: Assistants over private emails/docs using semantic (vector) + keyword retrieval, plus HyDE (hypothetical document embeddings) to improve recall.
- Rank Fusion & Reranking: Combining multiple retrieval strategies with Reciprocal Rank Fusion (RRF) and refining results via LLM-based reranking (with local fallback).
- Turkish NLP in the loop: Zemberek-backed normalization, stemming/morphology, spell-check, tokenization, and entity extraction to make search behave well in Turkish.
- Self-improving Product Matching: Learned-match cascade + hybrid search (vector + fuzzy) with score fusion for high-accuracy catalog matching.
Background
I studied Computer Science at Istanbul University and later earned an MBA from Sakarya University. Early in my career, I worked at Honda Türkiye and TIRSAN, where I learned the fundamentals of managing complex software projects and building relationships with stakeholders.
This site
This is my corner of the internet. I write here occasionally—in both English and Turkish—about technology, leadership, and whatever else is on my mind. No strict schedule, just sharing thoughts when something feels worth writing about.