About
20+ years driving IT strategy, product development, and digital transformation. Throughout my career as CIO, CTO, and DPO, I've led organizations through complex technical challenges — from building high-performing teams and optimizing processes to delivering products that align business goals with technical excellence.
My current research focus is LLM-native programming — designing languages and tooling where AI is a first-class author, not a code-completion assistant. Synoema is the result of that work: a language built from the ground up to align with how large language models tokenize, reason, and generate code.
End-to-end ownership of the software development lifecycle, from architecture decisions to release management. Proven ability to bridge the gap between business stakeholders and engineering teams, translating needs into actionable roadmaps.
Why Synoema
Every token an LLM generates costs time, money, and context window. Mainstream languages were designed for humans: verbose syntax, implicit context, irregular token boundaries. Synoema treats token efficiency and predictability as a core language property — not an afterthought.
The goal: LLMs that generate correct, verifiable Synoema code at a fraction of the cost of equivalent Python or TypeScript. Backed by measurable results — fine-tuned 3B models reaching 41%+ pass rate on benchmarks designed for 70B+ models.
If you're working on LLM code generation, AI tooling, or language design — I'd be glad to connect.