Description

I specialize in transforming complex, real-world challenges into cutting-edge, practical solutions. With a robust academic foundation, including peer-reviewed publications in advanced machine learning methodologies, I bridge the gap between innovative research and actionable business outcomes. My proven track record includes designing state-of-the-art AI systems, significantly enhancing operational efficiency, and unlocking entirely new customer acquisition channels. My expertise spans semantic search, retrieval-augmented generation, multimodal data understanding, automated document processing, and risk detection systems. Collaborating with industry leaders like Siemens and pioneering startups, I've consistently delivered tangible value, turning advanced AI concepts into scalable, production-ready tools that drive growth and strategic advantage.

Industry field of expertise

Languages

German
Native or bilingual
English
Fluent
Romanian
Native or bilingual

Workplace preferences

Remote only

Primarily works remotely

Keyspot AI
Senior Machine Learning Engineer
CONSULTING AND AUDITS
November 2024 - Today (1 year and 7 months)
Berlin, Germany
Developed an automated document revision detection tool for rapid identification of modifications in technical RFQ/RFP documents, greatly reducing manual verification time.
Created a multimodal model for precise data extraction from complex datasheets, streamlining the preparation of bids and proposals.
Implemented a hierarchical retrieval-augmented chatbot system for semantic understanding and accurate response generation, optimized to minimize false-positive retrieval.
Delivered automated risk detection algorithms that efficiently screen RFQ/RFP documents for compliance risks, enabling quick decision-making via an intuitive dashboard interface.
Siemens AG
Machine Learning Engineer
RESEARCH
April 2022 - October 2024 (2 years and 7 months)
Berlin, Germany
Engineered a retrieval-augmented generation pipeline incorporating hierarchical retrieval and context refinement for improved accuracy and reliability.
Created metrics for assessing context relevance and detecting hallucinations to enhance the quality and trustworthiness of AI outputs.
Developed novel unsupervised algorithms for semantic segmentation of text documents, supporting efficient content analysis (patent pending).
Built a multimodal encoder-decoder architecture integrating knowledge graphs for enhanced language modeling and data representation, reducing hallucinations in accuracy-critical business settings.
Designed a context-aware recommendation system for question generation tailored to technical documentation.
Siemens Advanta
Machine Learning Engineer
CONSULTING AND AUDITS
November 2020 - March 2022 (1 year and 5 months)
Munich, Germany
Built and deployed an extractive semantic search engine using advanced retrieval and reranking techniques.
Fine-tuned transformer-based machine learning models (TSDAE, MLM, CT, SimCSE) for precise domain adaptation on proprietary datasets.
Implemented comprehensive MLOps practices for efficient experiment management and deployment.
Developed a production-grade web application using FastAPI, Docker, and Angular, delivering intuitive user interaction aligned with corporate visual standards.