Work history

Five years of shipping AI that pays for itself.

Nov 2024 — Apr 2026

Budapest, HU

GenAI Lead & Software Engineer

Ominimo, Inc.

—Built Ominimo GenAI Studio, a FastAPI platform that generates production-ready chatbot backends from templates — cutting R&D from weeks to days per rollout.
—Designed and deployed insurance chatbots handling 60–70% of incoming customer requests end-to-end.
—Developed LLM-powered email pipelines ingesting Gmail via Pub/Sub & IMAP, classifying intents, extracting policy data, and routing tasks downstream.
—Implemented market-specific email agents (Hungary Adacta-flows, Netherlands GCP IMAP pipeline).
—Created a GenAI observability dashboard aggregating KPIs across markets: automation rate, intent accuracy, escalation rate, response times.

Aug 2024 — Nov 2024

Remote · US

Stealth

—Spearheaded a GenAI copilot for real-time Discounted Cash Flow models, enhancing stakeholder support.
—Designed a RAG pipeline with Guardrails to prevent jailbreaks; integrated Weave for LLM production evaluation, reaching 98% precision.
—Deployed a hybrid LLM + algorithmic model to standardize financial statements — 90% automation with S3-based data storage.

Jan 2024 — Jun 2024

Remote · US

Seeloz Inc.

—Applied DQN and PPO using StableBaselines3 and Ray RLLib for deep reinforcement learning research.
—Designed, integrated, and tested four novel features for the Seeloz Supply Chain Automation Suite AI Inference Engine.
—Conducted comprehensive code review and customer-data evaluation for leadership presentation.

May 2023 — Jan 2024

Remote · US

Cicero.ly

—Benchmarked LLMs using DeepEval and CoT techniques, reaching 95% response stability.
—Built production-ready ML pipeline for RAG-based synthesis with LangChain and FastAPI.
—Created a scheduled job summarizing and generating topics for 300+ new content posts daily with optimized latency.

Jul 2022 — May 2023

Sousse, TN

Cyber Division

—Built real-time scalable recommender systems for 200K+ products using hybrid models.
—Developed a RASA chatbot with CI/CD pipeline for NLP model maintenance.
—Used storytelling techniques to communicate AI feature impact to stakeholders.

Domain expertise

End-to-end design, template-based generation, production deployment, observability.

Intent classification, data extraction, automated replies, multi-market integration.

Document injection, guardrails, LLM evaluation, response precision optimization.

Real-time, scalable, hybrid (content + collaborative) — 200K+ product scale.

DQN, PPO using StableBaselines3 and Ray RLLib.

DCF modeling, statement standardization, hybrid LLM + algorithmic models.