
AI/ML Software Engineer and System Architect with 4+ years of experience designing scalable FinTech systems, time-series forecasting pipelines, and multi-agent architectures. Expert in end-to-end system design from architecture decision records (ADRs) to production deployment. Specialized in LLM-based financial agents, production RAG systems, and real-time inference using LangGraph, vLLM, and Triton Server. Proven track record: improved LLM accuracy from 80% to 95%; achieved 65–80% directional accuracy in crypto market prediction; reduced enterprise support costs by 30–50%. Experienced in large-scale model training (95K GPU hours on HPC). M.Sc. in AI candidate at FAU Germany. EU Blue Card holder with full-time work authorization.
AI/ML: Deep Learning, LLMs, RAG Systems, Multi-Agent Systems, GANs, NLP, Computer Vision, Time-Series Forecasting, Survival Analysis, Reinforcement Learning, Model Fine-Tuning, Model Evaluation, Real-Time Inference
LLM & Inference: LangGraph, LangChain, LangSmith, Langfuse, vLLM, Triton Server, AWS Bedrock, OpenRouter, Hugging Face Inference API, GPU Deployment (vastai)
ML Frameworks: PyTorch, Hugging Face Transformers, Scikit-learn, XGBoost, LightGBM, CatBoost, ONNX Runtime
Full-Stack: FastAPI, Flask, Django, PHP Laravel/Zend/Slim, React, Angular, Selenium
Javascript