Rand El-Shereef PhD PEng

• Technical lead for Advanced Analytics aims to transform analytical capabilities in chemicals and petrochemical manufacturing
• Partnered with industrial clients to identify opportunities for the application of PAT tools for business improvement
• Applied regression, and time-series modelling to manufacturing data to identify quality drift and improve process control.
• Optimized yield by building predictive models and dashboards for process parameters in partnership with operations and quality teams.
• Communicated new modelling insights and findings to experts and business leaders to promote data-driven decisions
• Promoted a strong quality mindset with a focus on data integrity, validation, and data governance
• Used process knowledge to support the data infrastructure team in developing, deploying and operating data pipelines
• Managed end-to-end data science projects, including data gathering and requirements specification, processing, analysis, ongoing deliverables, and presentations.
• Used a variety of data science tools, building and implementing models, using/creating algorithms, implementing solutions in production environments to solve different process-related use cases such as excursions detections, identification of key contributing factors towards low yield and troubleshooting .
• Provided operational and scientific support on multiple projects with a core role in delivering expertise in machine learning, modeling, and PAT implementation on real world manufacturing data

Key Responsibilities

• Collaborated with internal stakeholders from multiple departments to identify opportunity for applying data science to solve complex business challenges (i.e. maximize yield, process robustness, predictable supply, proactive identification of potential issues)
• Developed data workflows/data pipelines to extract, integrate data from multiple data sources (PI Historian and Discoverant) and multiple processes
• Developed/implemented algorithms using advanced statistical analysis and visualization of data
• Recommended implementation of Process Analytical Technologies to enable data capture for use in quantitative analysis and improved level of process understanding
  
  

Project 1: Process Comparability Analysis - Capacity Expansion Biologics (CEB)

• Developed workflows to collect and integrate data from (i) multiple processes (upstream processes, reactors, isolation, purification and final product) and (ii) from multiple data sources (PI Historian, Discoverant) in two drug manufacturing plants.
• Implemented a validation protocol to ensure data integrity, accuracy and completeness from source to final output
• Gained approval from multi-disciplinary CEB team, and executed data collection/ analysis methods following cGMP validation test functions.
• Delivered comparability reports to SMEs/ regulatory agencies to gain approval for drug manufacturing.

Project 2: Root Cause Analysis of Excursions in Product Quality

• Applied multivariate data analysis methods to relate polymerization structural patterns to process parameters.
• Identified the statistically significant contributors to excursions, communicated the outcomes to process SMEs, and confirmed the conclusions.
• Drafted and submitted the root cause analysis report to regulatory agencies.

Project 3: Statistical Equivalence Hypothesis Testing for New Instruments (pH and DO probes)

• Performed statistical equivalence/comparability tests between new and current instruments including: Determination of Means Allowable Difference (MAD) based on process knowledge, assessment of equivalency of precision and accuracy using Two One-sided t Tests (TOST)
• Implemented appropriate statistical methods to establish threshold difference acceptance criteria for comparability analysis.

Project 4: Online Viable Cell Density (VCD) Measurements in Commercial Scale Bioreactors

• Supported development, testing and validation of a correlation model to translate full spectrum sensor signal to inline VCD reading
• Introduced an online model health diagnostic tools for DCS implementation to check the reliability of every model based predicted VCD.
• Unreliable model predictions can be flagged to avoid making process decisions on poor quality predictions

Project 5: Prototype End-to-End Data-driven Framework of Manufacturing Commerical Process

• Supported development of prototype end-to-end data-driven framework of manufacturing process
• Developed mapping between product quality attributes and selected process parameters across process areas (including upstream and downstream processes)
• Defined aggregation rules for mixing steps across process areas
• Built prototype descriptive and predictive models for holistic process assessment to assess effect of upstream process parameters on downstream quality attributes

Project 6: Detection of Raw Material Variations using Vibrational Spectroscopy

• Collaborated with experts across Bayer sites to support investigation of using vibrational spectroscopy (Raman and NIR) to detect raw materials variations (for a fed-batch fermentation reactor)
• Developed a predictive MVDA model to predict final product yield of a fed-batch fermentation process from Raman measurements of raw materials, identified the best performing raw materials providing the highest yield
• Determined which combination of spectral preprocessing tools providing the best predictive model

Project 7: Detection of raw materials lot variations using FTIR

• Collaborated with experts across Bayer sites to support investigation of using FTIR to detect raw materials lot variations (for a commerical scale chemical reactor)
• demonstrated the potential for jointly evaluating following variations during batch progression: 1 )raw material lot-to-lot variations (e.g. occurrence of raw materials excursions)
  2) batch operational variations (e.g. occurrence of operational excursions)
  3) Association between raw materials and operational excursions

Project 1: Characterizing petrophysical properties using acoustic data and machine learning

• Investigated the relationship between acoustic properties and petrophysical properties for rocks in reservoirs located in Calgary, Alberta.
• Acoustic properties measured for rocks are correlated with their porosity variations, and permeability ranges.
• Machine learning was applied to predict textural properties and classify them based on (stiffness, porosity and crystalline structure from acoustic measurements.

Project 2: MSPC for online process monitoring of a large plant

• Implemented multivariate statistical process control (MSPC) for online process monitoring strategies, trends and shifts detection.
• MSPC reduced control charts numbers from 1300 SPC charts to 3 summary SPC charts capturing all types of variations.
• Resulted in improved efficiency, reduced work load and less paper work

Project 3: Monitoring and optimizing product quality attributes (2014)
Key Responsibilities

• Participated in R&D industrial projects aimed at demonstrating feasibility and robustness of multivariate approaches.
• Interacted with operation engineers at worldwide chemical companies to support defining optimization problems, specifying requirements and investigating options for solving problems.
• Supported development and validation of an acoustic-sensor for real-time monitoring of product quality attributes
• Conducted a Design of Experiment (DOE) followed by statistical MVDA-based approach for determining optimal operating conditions for large-scale production
• Conducted a statistical MVDA-based approach for determining optimal composition of raw materials required for maximizing product yield
• Identified combination of raw materials components contributing to high yield and maximum titer
• Provided recommendations to achieve optimal product quality.

Project 4: Perfusion/fed-batch production processes (2015)

• Supported design of an automated data aggregation platform for the management of information flow between the different stages of perfusion/fed-batch production processes and gathering the data into a format suitable for multivariate modeling
• Developed a statistical MVDA-based approach for perfusion/fed-batch production processes to (1) support, process monitoring, troubleshooting, root-cause analysis and identification of problematic raw material lots using SPC charts, (2) identify key parameters' effects on product quality and productivity at various stages throughout processes
  

• Played key role in developing data-based models from historical plant data for on-line monitoring, troubleshooting and optimization of chemical processes for improving product quality and reducing cost.
• Proposed a spectral preprocessing tool for a food processing company leading to improved prediction of product quality from NIR spectra.
• Supported development of soft sensors that predict difficult-to-measure product properties in real-time.

• Conducted basic and applied research on application of data-mining and chemometics of multi-wavelength fluorometry for monitoring membrane-based purification processes.
• Demonstrated that overall filtration performance can be inferred from fluorescence measurements acquired for feed, retentate and permeate streams during the progress of filtration.
• Developed a fluorescence spectroscopy-based inferential model for rapid, continuous prediction of feed, permeate and retentate compositions during the progress of filtration.
• Programmed computers to store, process and analyze data.
• Interpreted research findings and summarized data into reports.
• Attended industry conferences to broaden knowledge.
  

• Developed mathematical models of the thermoplastic structural foam molding process utilizing a chemical blowing agent
• Studied effects of processing parameters on the bubble growth dynamics such as processing temperature, concentration of blowing agent, and kinetics of decomposition reaction of blowing agent.
• Developed numerical solutions of bubble growth in Newtonian and Non-Newtonian polymer melts for isothermal conditions

• Supported introduction of new products and improving the reliability of existing products
• Reporting to Manager of Quality Assurance Group results and procedures of testing
• Supported quality engineering and other engineering groups with investigation of failure modes and failure mechanisms
• Conducted scratch durability testing and chemical durability testing on the products plastics frames, coatings to evaluate their mechanical robustness and to establish cosmetic durability standards for RIM\'s products
• Conducted accelerated life testing for the products to evaluate fatigue failures
• Involved with processing and analysis of data and developed standards for conducting analysis and presenting data Involved with modifying/maintaining quality standards for the products design and manufacturing