CTO / AI Systems Architect

I build typed, observable, auditable AI systems for regulated workflows.

Rust · Axum · Tokio · Postgres · OpenTelemetry · RAG · Agents · KYC/KYB/AML

AI prepares. Humans validate. Audit trails prove.

---

Systems I care about

typed workflows     > prompt spaghetti
domain models       > loose JSON blobs
audit events        > vague logs
evaluation          > vibes
human validation    > blind automation
observability       > guessing
permissions         > agent free-for-all
boring reliability  > demo theatre

Current surface area

Backend	Rust · Axum · Tokio · Postgres · SQL · Background Workers
AI Systems	RAG · Structured Outputs · Agents · LLM Orchestration · Evaluation
Ops	OpenTelemetry · Tracing · CI/CD · Docker · Cost / Latency Control
Domains	Fintech · KYC · KYB · AML · LCB-FT · Regulated Automation

Selected work

• Remolab — venture studio infrastructure for AI, fintech, and deep-tech products.
• Welcome Place — fintech and onboarding infrastructure for migrants and newcomers in Europe.
• CaseReady / RavenKYC — supervised AI for blocked KYC/KYB/AML/LCB-FT cases.
• AI Reading Club — foundational LLM papers, discussed from research to production.
• Rust AI Systems — ML, transformers, backend systems, and AI architecture in Rust.

Operating thesis

AI systems become useful in serious workflows only when they are:

  typed
  observable
  permissioned
  evaluated
  auditable
  boring enough to trust

Learning in public

I run the AI Reading Club : transformers, attention, BERT, generation, interpretability, scaling, fine-tuning, and alignment.

---