Rosely Peña

– Led analytics and Al experimentation strategy for customer support automation using Python, SQL, Snowflake and LLM-based chatbots, reducing support demand by 6,000+ tickets per month and generating ~$720K in annual operational savings. – Designed and evaluated Retrieval Augmented Generation (RAG) pipelines for Al customer support assistants, improving automated resolution rates by 30% and enabling scalable self-service analytics for operational teams. – Eliminated experimentation bottleneck by deploying an Al-powered MCP workflow integrated into Slack that automated experiment result analysis, driving a 200% increase in experiment throughput and removing all manual analysis overhead. – Built scalable analytics data models using dbt, Snowflake and Amazon Redshift, enabling reliable data pipelines and reducing reporting latency by 40%. – Led migration of Al chatbot platform in-house using LLM-based architectures, enabling 100% chat coverage and saving ~$360K annually in external vendor costs. – Spearheaded an Experimentation Training Program across the entire product team, spanning software engineers, product managers, and designers, empowering every member to independently launch and interpret A/B tests, eliminating the need for additional data science headcount to support experimentation at scale. – Conducted causal inference research using Propensity Score Matching to determine the direct effect of help-seeking on Pro retention, revealing a statistically significant 3.75% uplift for new trialing users, with an opportunity to add 390K Pro subscribers and $46.8M in incremental revenue by 2025. – Identified and sized a support deflection opportunity by analyzing 20K monthly uncategorized tickets, proposing an Al-powered deflection pathway using Al-powered chatbot projected to reduce ticket volume by 1,000 per month.

– Designed the analytical framework and technical definition for the end-to-end User Help Journey, mapping touchpoints across all help channels to enable predictive modeling at the session level and power friction diagnostics tied to churn and CSAT outcomes. – Built the first end-to-end analytics foundation for the Customer Support and Operations teams using SQL, Python and Snowflake, improving operational reporting coverage by 70% and enabling data-driven workforce planning. – Designed forecasting models to predict support demand and staffing requirements, improving workforce planning accuracy by 25%. – Developed operational dashboards using Looker for helpdesk teams and leadership, improving visibility of customer support KPIs and reducing decision response time by 35%. – Defined and standardized customer service metrics across the User Voice organization, establishing realistic performance targets. – Created a quantification framework for goal-setting and milestone tracking used in strategic planning sessions, enabling data-driven alignment across leadership and operational teams.

– Led data-driven operational analysis for packaging production lines, managing performance metrics across safety, quality, delivery, and cost KPIs for a production team of 15. – Applied Lean Six Sigma analytics and operational reporting to identify production inefficiencies, increasing operational efficiency by 10% across packaging processes. – Led loss elimination analytics initiatives that reduced packaging scrap waste by 12% and generated approximately $3,000 in monthly cost savings. – Built operational reporting frameworks and performance dashboards supporting daily production decision-making and resource planning.

– Designed and optimized manufacturing processes using extensive data analysis and applying Lean/Six Sigma methodologies for continuous improvement. – Led the successful Project Start-up of Joy 200 mL Conversion from Sachet to Pouch, achieving Instantaneous Start-up Results of 88% Process Reliability vs 85% target. – Created and implemented standardized work processes for Shift and Line Daily Direction Setting meetings across the site.