Elliott Luftman

Artificial Intelligence Machine Learning Intern @Cybernet Communications, Inc

Los Angeles, CA, US

Enhanced AI performance by fine-tuning large language models with internal datasets, improving accuracy and relevance in addressing customer issues- Streamlined customer support by integrating AI systems with internal databases, enabling real-time access to product information and customer service records- Developed automated workflows to connect AI with company systems, ensuring context-aware responses for faster and more effective issue resolution- Built an Enterprise-Grade AI Ticket Resolution System featuring Llama 3.3 70B integration, vector-based semantic search with sentence transformers, and real-time internet intelligence via SerpAPI - delivering 85%+ accuracy vs 40% baseline and reducing resolution times from 4+ hours to under 2 hours- Engineered Advanced Vector Similarity Search using FAISS indexing on historical tickets with domain-specific embeddings, replacing basic string matching with semantic understanding that finds conceptually similar issues with 95% faster search times (50ms vs 2-5 seconds)- Developed Self-Improving AI with Continuous Learning through comprehensive feedback collection, pattern recognition, and automated optimization - system learns from every technician interaction to enhance escalation accuracy from 72% to 90%+ and generates data-driven improvement recommendations

University of Southern California

Bachelor of Arts - BA, Intellegence and Cyber Operations

University of Southern California

Minor, Artificial Intelligence Applications