• Contributed to the development of Agentic AI and RAG for automated web application penetration testing. Using SFT, LoRa, MoE, FSDP. Libraries used: Langchain, Langsmith, Pydantic.

• Engineered a TLS Client Hello Parser in Rust.

• Created machine learning tools for anomaly detection, utilizing Transformer models and VAEs.

• Created concept for continuous security and DevSecOps best practices for the whole Org

• Enhanced Static and Dynamic Application Security Testing tools with Deep Learning

• Used Pytorch and contributed to Pytorch Geometric

• Researched Graph Neural Networks

• Researched Causal Inference for Deep Learning

• Research Interpretability and Explainability for ML models

• Resarched privacy-preserving LLMs

• Researched LLM-based Code Audits

• Compared Local Differentially Private and Central Differentially Private Deep Learning

• Conducted Membership and Inference Attacks against Machine Learning Models

• Tested privacy-preserving Federated Learning Models

• Developed an Extension for TensorFlow to support Federated Learning

• Developed Central and Local Differentially Private Federated Learning using TensorFlow