AI Strategy & De-risking
Board-ready clarity on where AI creates value — and where it creates exposure — before you commit budget, headcount, or tools.
- Roadmap
- Opportunity sizing
- Risk
- Vendor selection
AI builder & ISO/IEC 42001 auditor
I build AI systems — and I audit them.
I help leaders make AI decisions they won’t regret. Three years shipping production AI in legal, banking, and finance — and now auditing AI management systems against ISO/IEC 42001. I’ve shipped the thing, and I’ve audited the thing.
- Built & shipped
- Production AI, 3 yrs
- Audited
- ISO 42001 · 27001
- Based in
- Bucharest, RO
Shipped in regulated, high-stakes domains
Legal/Banking/Finance/KYC · AML/EU AI Act/ISO 42001/ISO 27001/MLOps
// The problem
You’re not confused about AI. You’re afraid of the wrong move.
The budget, the board, the regulator, the headline risk — the cost of a bad AI bet isn’t the bill, it’s the regret. What you actually want in the room is someone calm who has shipped real systems and knows exactly how they fail an audit.
BUILD
I’ve shipped the thing.
ML infrastructure at 500K+ docs/day, KYC/AML platforms for banks, document intelligence for legal teams. Real data, real latency, real adversaries.
AUDIT
I’ve audited the thing.
An active ISO/IEC 42001 and 27001 auditor assessing whether AI systems are governed, documented, and safe. Most people in AI have done one or the other.
// Services
From unclear ambition to a system you can defend.
Agentic Systems & AI Products
Production AI built for real workflows: LLM apps, RAG, agents, evaluation loops, and the application surface around them.
- LLM apps
- RAG
- Agents
- Evaluation
AI Governance & Compliance
EU AI Act risk classification and ISO/IEC 42001 readiness — from someone who audits these management systems for a living.
- EU AI Act
- ISO 42001
- ISO 27001
- Policy
ML Infrastructure & Security
Inference, pipelines, observability, and red-teaming for AI that has to survive real data, real latency, and real adversaries.
- Triton
- MLOps
- Observability
- Red team
// Selected work
Production AI in high-stakes, regulated domains.
CASE-01Finance
1,000+ articles/sec
High-throughput News Intelligence
NVIDIA Triton inference pipeline, distributed scraping, extraction, and monitoring — 500K+ docs/day at 99.9% uptime.
CASE-02Legal
RAG in production
Legal Document AI
Playbook ingestion, template parsing, and document-aware Q&A for legal workflows where a wrong answer is expensive.
CASE-03Banking
KYC/AML · multi-tenant
Compliance SaaS for Banks
Automated KYC/AML workflows serving multiple financial institutions, with graph data modeling and full observability.
CASE-04Assurance
ISO 42001 / 27001
AI Governance Audits
Assessing AI management systems and information security against certification requirements as an active auditor.
// Operating model
A practical path you can audit at every step.
The work isn’t model selection. It’s scoping the real decision, confronting data reality, designing the application, hardening it for production — and proving it’s governed.
- 01
Scope & de-risk
Define the decision the system must improve, and the failure cases you can’t afford. No model talk yet — just the real stakes.
- 02
Build the slice
Ship a narrow version on real data, with measurable outputs and a clear evaluation. Proof beats slideware.
- 03
Harden & govern
Add monitoring, latency control, and security — plus documentation that satisfies the EU AI Act and ISO 42001.
// Track record
- NOWCTO, Pythia Sociodynamical Technologies
- NOWAuditor, Cert Sud
- 2026ISO/IEC 42001:2023 Internal Auditor — TÜV Thüringen Karpat
- 2025AI Engineer, CogniSync
- 2023ISO/IEC 27001:2022 External Auditor — CERTINSPECT Register
- 2022Co-Founder & CTO, Pythia Sociodynamical Technologies
- 2020IT Administrator, Cert Sud SRL
- 2019Research & Development Intern, Continental Powertrain Engineering
// Stack
DSPyPyTorchScikit-LearnNVIDIA TritonNLPCVDLPythonFastAPINestJSLitestarRESTful APIsPandasMongoDBPostgreSQLRedisRabbitMQSurrealDBTypescriptNext.jsReactRemixTailwindCSSDockerAWSAzureCI/CDPrometheusGrafanaNixRomanianEnglishGerman
// Start here
Make the AI decision you won’t regret.
Bring a decision, a bottleneck, or a system you need to trust. Thirty minutes, no slides — we’ll find out fast whether I can help.