Zerub Roberts Email & Phone Number

Melbourne, Victoria, Australia

● Provided consulting services to clients across multiple sectors (engineering, utilities, university, water, and technology); identified business challenges and needs and determined how the use of predictive or prescriptive analytics could be leveraged by the business.● Experienced in connecting to, interrogating and integrating a wide range of data sources - legacy, finance and ERP systems, REST API connectors and databases (MS Access, SQL Server, AWS Redshift, Hadoop) and IoT Sensor data.● Built end-to-end Data/BI Solutions by integrating various data sources (SQL, legacy systems, APIs), performing data modeling, creating dashboards for different user personas, and including advanced analytical capabilities (ML).● Piloted sophisticated advanced & innovative analytics solutions and demos to prove value for clients: 💡 Geo-analytics and route optimization using ArcGIS and Google Maps API 💡 Predictive asset maintenance (ensemble models using Python integration with Qlik - Pytools library) 💡 Predicting Weather-related power failures (this solution is now implemented across Eastern Victoria Power Maintenance company) 💡 Detecting lack of protective clothing on construction sites using OpenCV, AWS Sagemaker, Rekognition, and real-time SMS alerts.

United States

Instructor/project evaluator for the Tableau Data Analytics Program offered by Pathstream to various Universities and adult learners at Amazon & Walmart. My responsibilities include - ● Assessing student project submissions (submitted via the Pathstream platform) against a rubric to determine an overall score, within 72 hrs of project submission.● Providing constructive and actionable qualitative feedback that leads to deeper understanding and skill building, using a professional yet friendly, growth-mindset-oriented tone.● Professionally and promptly responding to student questions related to their score, in coordination with the instructor for the course.● Providing feedback to the instructional team & learning experience design leads to improve curriculum and projects.● Developing project-specific comment bank, highlighting common student errors to enable easier onboarding for new project evaluators.

Australia

Australia

United Kingdom

𝗞𝗲𝘆 𝗔𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵𝗺𝗲𝗻𝘁🚫 Problem: CS agents are a big chunk of operational costs. Need automated email triaging due to increased inbound volume while onboarding bigger clients.💡 Solution: Built & deployed a Text Classifier model on AWS Sagemaker to deflect 40% of inbound emails comms & with a weighted accuracy of 70%🎯Outcome: Deflected 20% of new inbound emails per day (about 1200 emails) & sent auto-responses, saving ~100hrs/day (AHT of 5 mins/email).🔹Deployed an inbound comms volume forecasting model (across multiple channels & countries), with a MAPE of 74% to enable CS teams to better anticipate periods of peak demands & roster agents accordingly to lower customer wait times.🔹Quantified the $-value impact of various ML models on the business and created dashboards to monitor model performance and track real-time cost savings.🔹Followed best practices in MLOps (experiment tracking with MLFlow, model/data versioning, monitoring model drift, CI/CD pipelines, etc.)

United Kingdom

Took ownership and worked with cross-functional teams to implement a suite of internal & client-facing dashboards, built key metrics, and automated reporting -🔸Finance: Client Profitability Dashboard. Associated all costs (stripe transactions, customer service time spent, fees, messaging costs) & profits per each client.🔸Customer Service: Built a comprehensive Agent Scorecard dashboard tracking all aspects of 100+ CS agents (productivity, efficiency, compliance, and knowledge checks). Quantified metrics around Avg. Handle Time, Response Time, and Resolution Time and created a unified score for ranking purposes.🔸Client-facing dashboards - Moved client reporting from monthly PDF reports by implementing a Tableau server with row-level-security features, providing real-time insights into their customers.🔸Product & Data: Collaborated with Product & Data Science to track the impact of various ML models on different engagement metrics (click rate, conversion rate, with the entire funnel visualization).

Melbourne, Victoria, Australia

● Learning advisor for the Graduate Data Analytics Certificate offered to students at Queensland University of Technology (QUT Online).● Responsible for facilitating the planned academic content and acting as the first point of contact and support for our online students.● Marking, processing and submitting assessment results according to established guidelines within requested timeframes.● Being visible and available to students, maintaining a presence on discussion boards every 24-48 hours.

Melbourne, Victoria, Australia

💡𝗞𝗲𝘆 𝗔𝗰𝗰𝗼𝗺𝗽𝗹𝗶𝘀𝗵𝗺𝗲𝗻𝘁 Created custom visualizations for the research team to visualize complicated interactions within large datasets using UpSetPlot in Python (for plotting a 24-set intersection graph) & non-native Tableau visualizations (sankey, radar, custom-drawn)● Delivered data analytics and reporting services to headspace centers, Primary Health Networks (PHNs), and internal stakeholders● Designed, developed, and standardized tableau dashboards for the Early Psychosis Program (hYEPP) which is now used by center managers and service providers across 130 centers throughout Australia.● Integrated ABS data with internal data sources to create statistical models for determining the optimal locations for opening new headspace centers using Python and ArcGIS.● Developed python scripts to automate monthly and quarterly data extracts for compliance and reporting, data migration quality testing, etc.,Tech Stack: Tableau, Python, MS SQL Server

Master Of Data Analytics, Data Science

Microsoft Azure Machine Learning

Bachelor Of Computer Science, Mathematics And Computer Science