Joseph S. Naraguma, PhD

512-***-****

ad2tsi@r.postjobfree.com

Professional Experience

Paloaltonetworks Inc. Principal Data Scientist 2022 -Present As part of the Data Science team, I lead analytics team supporting Growth Marketing operations within CMO, across product lines especially the 6 company flagship products (Firewall, Cortex, CDSS, Unit 42, Prisma Cloud and SASE) and product stakeholders, including performance analytics around those products. Modeling activities focus on the propensity to buy and lifetime value (LTV) across those products, but also assessing customer engagement metrics.

Dataiku DSS –integrated with Python for special feature engineering libraries - is the platform of choice for the Data Science practice for both modeling design, deployment, and automation. We also leverage both GCP for storage – and custom SQL syntax for ingestion and BigQuery for exploration. Often leverage ChatGPT to analyze and explain technical support use cases and come up with product specific and high performing and reusable prompts

HCL Americas Inc. Sr. Data Scientist 2019 -2022

In this role, I was typically engaged in a specific project with a paying client; currently. Typically, I have teamed up a data engineer, and MERN developer to understand the client ticket system data in different ERPs to determine platform reliability issues, identify incident root cause and recommend how to eliminate tickets, and wherever possible build an automated process to channel incidents to right channel for resolution and ultimately suggest process improvement areas. The critical piece of the project was precisely where my team read enterprise data from GCP, define and manipulate data through SQL and applied NLP approach, leveraging key python libraries ( nltk, spacy, gensim, pattern3, etc.) to ticket classification and automation through topic modeling across their ERPs. Preprocessing the unstructured data is essential and includes steps such as eliminating useless words and characters, tokenization, and stemming, etc.

This is an example where Machine Learning is badly needed. An incident may be composed of multiple topics. Users can indeed complain of several things. Typically, only one of the topics is dominant. Key to preparing for topic modeling is precisely to extract dominant topic for each sentence and shows the weight of the topic and associated keywords, an area where my team leverages business SME inputs to come up with actionable labels. In my role as SME in Analytics, constant collaboration with the stakeholders/business is essential. Prior to this, I lead the ML Advertisement Insertion initiative within the client’s Ad Inventory Management organization on behalf of HCL Americas. The net impact of the project has been to improve advertisement revenue by finding the best time to play ads in the more than 150 networks and optimize the resources available in term of slots available to play those ads. A combination of seasonal ARIMAX and Holt Winters exponential smoothing models have been customized –using mainly python prophet and other specialized libraries- to each of more than 150 networks they target in the different video hub offices (VHO)/markets. Managing a team of 3, we were able to deliver a framework to forecast and guide networks to target the most profitable times of the day Dell Services/NTT Data Services Sr. Data Scientist 2015 - 2019 As part of the Advanced Analytics team under Business Intelligence & Analytics Practice, I lead analytics support on a key account (Ascension Health Network) providing analytical methodology support/ design, knowledge transfer, and insights generation to the business. I have leveraged several data sources (e.g., CRM data, web interactions and social listening of patients/customers) to see how engaged users were to the different LOB websites and applied both supervised learning -classification and regression- and unsupervised learning, NLP/text mining, including word clouds/embedding, sentiment analysis, and topic modeling. Periodically refresh their models as new data comes in. SQL has always been a critical piece of my data definition and manipulation work before modeling/analysis

Whether dealing with fulltime engagements or POCs, the typical process I follow, in collaboration with stakeholders/clients, leading to a statement of work (SOW) agreed by both parties, goes from defining the business objective, devise an appropriate analytical approach, secure data sources all way in different environments (On premises and/or clouds) to model development and deployment of the analytic solution, and final presentation to the client, documentation and knowledge upon client request. My interaction with the project lead/stakeholder ought to be constant throughout the project

Contract Consulting. Sr. Analytical Consultant 2012 –2015 Consulting for several clients in different industries (mainly professional services, financial services). Had lead roles in the clients’ Decision Science teams. Worked on various projects for clients in services/financial/insurance

/services industries, supervising and mentoring other analysts in the team. Projects have ranged from assessing profitability prospecting of customer databases (internal and external, e.g., D&B), assessing different media investment to sales through market mix modeling (MMM), evaluated predictive factors on the efficiency of call centers, stress testing for banking segment of the business, evaluating macro-economic factors on new/lost membership and product breadth/deepening, auditing and modeling credit card data using mainly a combination of both data mining and forecasting SAS tools, sometimes Hive/Hadoop for parallel data processing. SAS Institute Inc. Sr Analytical Consultant 2011 – 2012 Responsibilities included delivering statistical models and forecasts to clients in various industries. Relevant activities include problem definition, requirements gathering, project scoping (estimation, defining deliverables and evaluation metrics), proposal development, and advanced analytics and delivery of insight and presentations to client management. Supported projects in the areas of Healthcare, Retail and Manufacturing, Financial Services, Education, Content and Entertainment, Telecommunications, Advertising, CPG, and Transportation. Cisco Systems Inc. Sr Marketing Analytics Consultant 2008 - 2011 Data Mining consultant role: primarily developed propensity-to-respond models from Cisco (and Cisco Partners), marketing and customer-facing interactions (Events, Seminars, Subscriptions) leveraging contact attributes, frequency, volume of interactions, along with firmographics and bookings derived variables, in support of marketing campaigns and Market Intelligence organization

Forecasting role (Year 2008 thru May 2010): as part of the Strategic Marketing Organization, forecast Cisco bookings in support of sales and marketing across geographies, anticipating future levels, for both on service and product bookings based on historical records and the changing economic conditions

• Access large customer databases extract relevant booking fields and build working files to Exploratory Data Analysis and statistical modeling.

• Leverage Global Insight economic indicators, appropriate transformations on them (reduction techniques) and gage their impact on future forecasts using a combination of SAS/ETS, SAS/HPF, and SAS/Forecast Studio tools

• Identify business strategies to support internal organizations (Operations and Planning, Strategy) and present a data driven and more accurate view than the sales force forecasts

• Update/revamp forecast models on monthly and quarterly basis, adjusting for seasonality whenever possible, and creating key accuracy metrics for the different audiences

• Assess past impact and predict the future impact of various marketing tactics on sales. SAS Institute Inc. Analytical Consultant 2006 - 2008 Responsible for the sales support and promotion of SAS Data Mining and Analytical products through pre-sale training and delivery. Created standard demonstrations of the products (Enterprise Miner, SAS/STAT) for proof of concepts that created and closed several opportunities across industries. Prepared winning proposals and delivered business outputs to various clients.

• Wrote proposals and SOW to support SAS solutions on client sites across industries, from data procurement, EDA (exploratory data analysis), modeling, validation/scoring and deployment

• Leverage clustering, classification techniques (decision trees, neural networks), and MMM to improve targeting and acquisition of a variety of client databases across industries; areas of concentration included propensity to respond to insurance offerings, media impact, card membership, student enrollment etc.

• Developed models as base for demonstrations to customers or groups of customers to showcase how quickly the tools can build a model and deploy them without having to write new code or redesign new diagram flow. The effort reduced the time to build models by 50%.

• Implemented model comparison features of EM to select the best performing models among several algorithms (decision trees, neural networks, regression, etc.) but also leverage their Text Mining node for vocabulary-based vectorization to enhance its predictive analytics capability Education

PhD, Applied Statistics, University of Arkansas, Fayetteville, AR BS, General Sciences, University of Burundi, East Africa Data sources

Familiarity with data warehouse concepts, ETL, and data integration. Have leveraged Google Analytics and several RDBMS including MS SQL, Teradata, IBM DB2, Oracle/Oracle PL/SQL, GCP, Microsoft Access, and experience with NoSQL databases as enabled by big-data frameworks (e.g., Hadoop, Spark) Statistical and Analytic Consulting skills

Extensive experience with Predictive analytics’ algorithms including GLM, Survival analysis/Predictive maintenance, Logistic Regression, Neural Networks, Nonparametric methods, Multivariate Methods, Marketing Mix modeling

(MMM), Classification modeling- Decision Trees/KNN, SVM, Random Forests, Factor analysis, Cluster analysis/K- Means, CHAID, Two-stage regression, Multidimensional Scaling, Simulation, Scenario analysis, A/B testing, Data/text mining, and Forecasting

Market Research and Analytics Consulting Skills

Marketing Research Methods (Marketing Research Institute certified); Loyalty Research, Client Facing Experience across industries, familiarity with agile methodology for collaboration, Project management, experience with report writing and ability to communicate results from complex statistical/econometric techniques in a non-technical format FCRA certified

Software Skills/Packages

Proficiency in SQL, R, Python & PySpark/Jupiter notebooks, Tibco/Statistica, SPSS, Tableau, Google Analytics, Knime, Dataiku DSS (Certified), Alteryx, in addition to a SAS professional certification and experience with Hive/Hadoop, AWS, GCP, MS Azure ML Studio, on both Windows and Unix environments Language Skills

English & French