Mohamed Elashry

NLP Data Scientist II at WideBot

Led a dedicated AI team delivering 45+ customized LLM chatbots across finance, healthcare, HR, legal, and support domains; developed Arabic semantic search and AQL-GPT chatbot with dialect support, boosting query precision by 30%. Built a video transcription pipeline for Arabic dialects with time-aligned outputs (15% accuracy improvement) and launched AI-powered digital commerce solutions to enhance customer engagement. Implemented auto-scaling for chatbot deployment and integrated real-time analytics for performance monitoring.

Led a cross-functional team of AI engineers to design and deploy production-grade LLM-driven systems at scale for finance, healthcare, HR, legal, and support domains. Built multilingual AI solutions and RAG-based retrieval pipelines; delivered 45+ customized LLM chatbots; defined end-to-end pipelines, governance, and monitoring. Collaborated with product, data, and engineering partners to align AI capabilities with business goals, improving uptime, reliability, and decision-making in enterprise environments.

NLP Data Scientist II at WideBot

Developed Arabic semantic search system and AQL-GPT chatbot with dialect support, boosting query precision by 30%. Constructed production-grade AI pipelines across 5+ enterprise clients.

Senior NLP Data Scientist & Team Lead at 3i-vision

Led a team of 6 AI engineers, delivering 45+ customized LLM chatbots for finance, healthcare, HR, legal, and support domains. Built production-grade AI pipelines for Arabic and English, processed 100+ document formats, and boosted accuracy to enable faster decision-making across enterprise use cases.

NLP Data Scientist II at WideBot

Developed Arabic semantic search system and AQL-GPT chatbot with dialect support, boosting query precision by 30%. Constructed a video transcription pipeline supporting Arabic dialects with time-aligned outputs, increasing transcription accuracy by 15%. Launched Hulul, an AI-driven digital commerce solution, increasing customer engagement by 25% for 50+ MEN Arabic dialects. Fine-tuned large-scale language models for domain-specific use cases. Established and implemented auto-scaling infrastructure for chatbot deployment, enhancing uptime and reliability. Integrated real-time analytics for chatbot performance, providing actionable insights and elevating response quality. Crafted multilingual NLP pipelines supporting over 5 MEN Arabic dialects, driving user engagement by 20%. Released an A/B testing framework to assess chatbot performance, increasing conversion rates by 10%. Devised cash collection prediction systems for TMGG Group, forecasting inflows with 85% accuracy. Deployed predict

Senior NLP Data Scientist & Team Lead at 3I-vision

Led and mentored a team of 6 AI engineers, delivering 45+ customized LLM chatbots across finance, healthcare, HR, legal, and support domains. Developed a multi-agent chatbot stack using LangChain and LangGraph, enriching user interaction by 40%. Designed a production-grade document AI platform (OCR, classification, summarization) for Arabic and English, processing 100+ document formats and boosting accuracy. Built advanced LLM pipelines using SetFit and Comprehend IT, along with a synthetic OCR dataset containing 56+ Arabic fonts. Launched AI models across cloud and edge environments using ONNX and OpenVINO for 15+ enterprise clients, reducing latency by 30% and accelerating deployment timelines. Engineered enterprise-ready Q&A systems with GraphRAG and quantized LLaMA3.1 (70B) via llama.cpp, significantly optimizing inference speed and resource efficiency. Directed the design of custom AI models to automated document processing workflows in 10+ industry sectors, saving clients 1,000+

NLP Data Scientist II at SAP

Integrated AI solutions with SAP ERP across 5+ finance departments; automated cash collection and bank document processing using OCR/NER, cutting manual processing time by up to 75%. Designed a real estate Q&A chatbot handling 500+ queries weekly and developed cash collection prediction models with 85% accuracy. Enhanced existing AI models for financial document processing and streamlined reporting, saving hundreds of hours monthly; deployed a personalized real estate recommendations engine.

NLP Data Scientist I at WideBot

Implemented NLP/NLU pipelines for Arabic language processing, enabling improved user interactions and enterprise-scale deployments across multiple domains.

NLP Data Scientist I at CIC & SAP

Integrated AI solutions with SAP ERP systems, enabling workflow automation across five financial departments. Automated Intelligent Bank Reconciliation (IBR) to process bank documents using OCR and NER, cutting manual processing time by 75%. Designed a multilingual Q&A chatbot supporting 500+ customer queries weekly, increasing conversion rates by 35%. Developed multilingual NLP pipelines supporting over 5 Arabic dialects, driving user engagement by 20%. Revised and updated automated A/B testing framework to assess chatbot performance, enhancing conversion rates by 10%. Led cross-department workshops to integrate AI systems into SAP workflows, raising adoption rates by 40%.

Data Scientist & ML Engineer at Methanex

Directed data migration strategies, including data cleaning and validation for secure, compliant migrations; improved data accuracy. Designed AI-assisted QA and reporting components to support finance and operations domains.