Mohamed Rashad
Nlp Ml Research Engineer @WitnessAI
Signup · Get unlimited contacts
WORK HISTORY
Nlp Ml Research Engineer @WitnessAI
Mountain View, CA, US
Working on enhancing hierarchical clustering for better topic detection and user behavior prediction- Working on refining RL-based agents to increase filtering system adaptability and precision in real-time applications- Fine-tuned small LLMs using RL techniques (GRPO, ORPO, DPO) for real-time sync filtering, achieving response times of <180 ms on A10 GPU hardware- Developed a policy-driven content moderation system to block disallowed topics and ensure chatbot compliance with business policies- Created an agentic knowledge graph by clustering user prompts with hierarchical clustering, providing admins with deep insights into user interactions and chatbot performance- Optimized hierarchical clustering with proxy energy-based functions via agentic LLMs, improving topic categorization, search efficiency, and real-time performance.
EDUCATION
Ain Shams University
Bachelor's degree
ABOUT MOHAMED RASHAD
I\'m a passionate and results-driven NLP Research Engineer with a keen interest in developing cutting-edge solutions that leverage the power of language to unlock valuable insights from data. I specialize in multilingual text processing, information summarization, and entity analytics. My work includes developing innovative NLP models like synonym discovery, relation extraction, and graph-based/query-based summarization. I\'m also focused on advancing document understanding through computer vision models and parser-pipeline modules. With a background in NLP R&D, I\'ve contributed to unsupervised probabilistic frameworks and created specialized language models for regional text classification. As a Teaching Assistant, I mentored and supported students in NLP and Computer Vision, fostering their AI skills. Let\'s connect and explore exciting opportunities in NLP and AI together!
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.