Mohamed Derri Email & Phone Number

Dallas, Texas, United States

• Led the development and training of custom AI models for clients such as OpenAI and Google through reinforcement learning with human feedback, achieving improvement in model performance and accuracy.
• Led prompt engineering initiatives, ensuring the delivery of concise and effective prompts to large language models (LLMs), resulting in enhanced model comprehension and performance.
• Implemented reinforcement learning techniques to facilitate LLM training optimization, leveraging human feedback to iteratively improve model capabilities and accuracy.
• Mentored and audited the work of a team of 50 advanced AI coders, ensuring compliance with all relevant regulations and guidelines.
• Demonstrated dedication by working overtime to ensure timely delivery of projects, reducing project completion time.
• Implemented innovative techniques for fast review and audit processes, reducing turnaround time and enabling more agile development cycles.

Dallas, Texas, United States

• Trained and refined state-of-the-art Large Language Models developed by industry leaders like OpenAI, Google, and Meta, achieving notable improvements in model performance and accuracy through reinforcement learning.
• Spearheaded prompt engineering initiatives, ensuring the delivery of concise and effective prompts to large language models (LLMs), enhancing model comprehension and performance and safety.
• Implemented innovative reinforcement learning techniques to optimize LLM training processes, leveraging human feedback to iteratively enhance model capabilities and accuracy.
• Critically evaluated and corrected responses generated by Large Language Models, ensuring their accuracy and relevance for various applications.
• Assessed multiple responses from models for the same prompt, selecting the most appropriate and correct the others to enhance model performance.
• Mentored, trained and audited a team of 50 advanced AI coders, ensuring compliance with all relevant regulations and guidelines while fostering a culture of excellence and continuous improvement.

Fairfield, Iowa, United States

• Conducting research in deep learning with a specific emphasis on the next gen of Scene Graph Generation SGG.
• Exploring and implementing state-of-the-art models and algorithms in deep learning for image analysis.
• Contributing to academic publications and presentations, sharing findings with the research communityTechnologies: Python, TensorFlow, PyTorch, CNN, R-CNN, SSG

Zurich, Switzerland

• Leveraged graph structures to enhance accuracy of (ER) for detecting money laundering and compliance violations.
• Pioneered the integration of graph-based machine learning techniques to refine entity resolution (ER) processes, resulting in a 12% reduction in false positives.
• Engineered a multithreaded entity resolution engine using Python, Java, and OraclePGX, delivering a 20% increase in efficiency.
• Orchestrated rule-based algorithms and machine learning strategies to enhance ER scoring accuracy by 15%.
• Integrated Google Geocode API to enrich entity data, augmenting data quality and strengthening ER precisionby 7%.Technologies: Graph ML, python, recordlinkage, google geocode api, java, data studio, pandas, numpy.

• Conducted extensive exploratory data analysis on complex psychological datasets, uncovering critical patterns.
• built a machine learning model that allows processing and analysis of psychological tests in order to predict pathologies achieving a 92% accuracy rate.
• Designed and implemented a RESTful API using Python Flask delivering ML solutions, seamlessly integratingSQLAlchemy and Alembic for agile database management and migration.
• Designed and implemented a data visualization dashboard using PowerBI, enabling stakeholders to monitorkey performance metrics in real-time.
• CollaboratedwithdatascientiststoimplementXGBoostalgorithms,reducingpredictionerrorsby25%.Technologies: python, flask, MySQL/SQLAlchemy, Alembic, PowerBI, Supervised/Unsupervised ML, pandas

Casablanca Et Périphérie

• Led development of a cutting-edge property valuation platform, yielding 15% higher accuracy compared to traditional methods.
• Collaborated with data scientists to implement advanced ML algorithms, reducing prediction errors by 20% using regression and gradient boosting techniques.
• Conducted rigorous testing and validation, achieving a 95% confidence level in valuation predictions.
• Implemented interactive data visualizations, leading to a 50% increase in user comprehension of property trends and valuation drivers.
• Contributed to technical discussions, providing key insights that guided model selection and project direction.Technologies: Python, PowerBI, Pandas, TensorFlow, Java/spring boot, MySQL, Supervised/Unsupervised ML.

Casablanca Metropolitan Area

Master'S Degree, Computer Science

Diplôme D'Ingénieur, Software Engineer, 18.5/20

Architecte En Technologie Du Numérique, Computer Science

Classes Préparatoires Intégrées

Activities and Societies: Membre au Club ENACTUS ENSA KhouribgaLes années Préparatoires Intégrées proposent aux étudiants une formation.

Baccalauréat Internationale En Sciences Mathématiques B Option Française