Omar Kamal Hosney
Full-time · Santa Clara County, California, United States · On-site.
Full-time · Santa Clara County, California, United States · Remote.
Contract · Santa Clara County, California, United States · Remote. Energy-sector optimization and intelligent agent systems.
- Architected multi-agent AI systems using Cerebras and Groq, achieving 2,400 tokens/sec for autonomous decision-making
- Developed a Retrieval-Augmented Generation (RAG) system enabling LLMs to access real-time data for enhanced tool orchestration
- Fine-tuned a Qwen3-8B LLM for SQL generation at 95% accuracy, enabling natural-language-to-SQL translation for energy-sector databases
Full-time · 1 yr 1 mo · Hybrid, United States. Role progression: Senior Data Scientist (Jan 2024 – Dec 2024) → Senior Staff Data Scientist (Dec 2024 – Jan 2025). Led the development of advanced AI-driven systems and frameworks focused on personalized recommendations, cost-effectiveness, and scalability.
- Developed a sophisticated personalized collection framework using embedding and multi-tower models, balancing effectiveness and cost efficiency
- Created and implemented comprehensive documentation and design practices that became team standards
- Built an innovative demo tool for showcasing AI frameworks, subsequently adopted by the product team for specification and documentation
- Led technical presentations to leadership, communicating complex AI concepts and trade-offs
- Deep Learning: PyTorch, TensorFlow, Neural Networks, Multi-Tower Architecture
- AI/ML: Generative AI, Large Language Models (LLMs), Vector Embeddings, Semantic Search
Oversaw fintech data strategy combining AI-driven underwriting and advanced analytics. Developed insights from diverse data sources, constructed scientific data models, and designed statistical and data-mining products. Led initiatives to design and execute experiments that optimize data-driven business models.
- Large Language Model implementation
- Machine Learning systems development
- Data Science leadership
Directed data analytics for a stealth startup focused on subscription revenue optimization. Spearheaded development of a patent-pending algorithm for optimal business configuration. Introduced the innovative "Pi" metric for improvement areas.
- Python and NumPy development
- Algorithm design
- Statistical analysis
Spearheaded the integration of data analytics and led the formation of a dedicated data science team. Implemented comprehensive data analysis solutions across departments.
- R programming
- Statistical Analysis
- Churn Analysis
- Tableau visualization
Led software testing and quality assurance initiatives. Developed testing tools and analytics dashboards.
- Software Testing
- Quality Management
- Six Sigma Implementation
Developed automotive embedded software and validation systems.
Led the Class 5 Switch Project testing team.
Developed software solutions and led Quality Management System deployment.
Conducted technical market research and business development.
Worked on GSM mobile networks and developed embedded applications.
Product Recommender with Contextual Data
2024Advanced neural-network models built with PyTorch for personalized recommendation systems.
ChatGPT-4 Diagram Generation
2023Prompt designs enabling ChatGPT-4 to generate diverse diagrams via Mermaid, D3, and PlantUML.
Child-Centric Conversational AI
2023A Streamlit chatbot engineered with LangChain and OpenAI's GPT-4.
Cash CoPilot
2023Integrated ChatGPT with the Mexican Tax Authority for real-time financial analysis.
Customer Spending Alert System
2022Statistical process-control system for financial monitoring.
Bank Statement Parser
2021Parsing tool for Mexican bank statements with advanced security features.
Startup-VC Matching Algorithm
2017Matching system using word embeddings and cosine similarity.
Master of Business Administration
Diploma — Computer Networking & Software Programming
Bachelor — Electrical & Electronics Engineering
Patent (Filed)
Apparatus and method for predicting future incremental churn from a recurring-revenue product.
Advanced GraphRAG & Knowledge Agents: A Friendly, Illustrated Guide to Graph-Based Retrieval
A friendly, illustrated guide to graph-based retrieval, GraphRAG, and knowledge agents.
Diagrams as Code: Mermaid, PlantUML, D2 & AI-Generated Diagrams
Book exploring diagram-as-code tools and LLM-driven diagram generation.
Technical Blog on Medium
Active technical writer publishing tutorials on AI technologies.