Mike Erlihson

• Leading the development of AI and ML-based components for DLP (data loss prevention) agents
• Fine-tuning small on-device language models for NER, classification and reasoning tasks

• Developed ML-based anomaly detection techniques for API attack detection
• Created and deployed ML solutions for API structure learning

• Conducted algorithmic research for the Pipl search engine, specializing in record linkage
• Developed advanced NER solutions and unsupervised clustering methods on graphs

• Developed algorithms for sentiment, topic modeling and relevance analysis of text
• Created solutions for topic modeling and deep learning-based snippet extraction

• Developed deep neural network-based ML algorithms for smart flash memory controllers
• Built software architectures for Flash Controllers

• Taught the course “OFDM-based Communication Systems for graduate students.”

Was in charge of development of advanced OFDMA-based signal processing algorithm for base stations and edge devices