Portfolio
Agentic Workflows
- Open Knowledge Format (OKF) Enablement Engine A stateful, agentic pipeline for continuous revenue enablement that treats institutional knowledge as a dynamic data graph rather than static documentation. By mapping individual personas to modular “Open Knowledge” playbooks, the engine uses local LLM orchestration to synthesize hyper-personalized coaching insights. The system includes an automated feedback loop where manager-reviewed reports are ingested to update learner skill profiles, creating an evolving training record that tracks competency progression over time. Based on Google’s Open Knowledge Format.
- Q2C Policy Interceptor A LangGraph-based guardrail agent that intercepts deal parameters in the Quote-to-Cash flow, validates them against configurable discount and term policies, and returns a structured APPROVED/FLAGGED verdict with a rep-facing next-best-action, cutting off Deal Desk escalation loops before they start.
- Revenue Enablement Content Generator A local Revenue Enablement content generator that analyzes a technical specifications document (similar to what would be found in Engineering or Product documentation in Jira) and creates a product/feature training guide suitable for a Sales team including the feature brief, value, objection handling, and a brief quiz.
- Multi-Agent Support Router (MASR) A localized triage engine that ingests incoming customer tickets, extracts intent markers using local LLM JSON schema enforcement, queries on-premises vector indexes, and synthesizes a context-aware response draft entirely behind your firewall.
- Automated Sales Playbook Generator An intelligent GTM automation pipeline that joins structured firmographic metrics with raw text dossiers to identify high-value pipeline signals—specifically flagging newly funded target companies where an executive is a documented former platform user—and outputs a tailormade outbound sequence.
Browser Tools
Interactive Simulation
- B2B Data Hygiene Simulator: An interactive experience to learn about fundamental data quality principles including deduplication, format standardization, and enrichment. These are essential concepts for data sellers and others involved in GTM roles.
- Marketing Mix Model Explorer: An interactive marketing mix model simulation using Hill-curve saturation and market-driven model interaction features. Built with R and Shiny.
- Data Decay & Revenue Leakage Analytics: Based on industry-standard rates (Gartner) for annual B2B data decay and revenue impact. Built with R and Shiny.
- ABO Blood Type Interaction: Based on previous work where I co-created original curriculum and 3D printed models that were designed to help learners understand ABO blood type compatibility. The core interaction of that original curriculum is recreated here in an online medium while the original was intended as a physical model for hands-on exploration.
- Firefly Meadow: Uses published data to simulate a meadow containing seven firefly species. Users can select the species to populate the meadow and select individual fireflies to activate a visualization and a sonification of the underlying data.
AI-Powered Technical Enablement Samples
These interactive examples were created with RStudio, Quarto, Gemini, and Claude.
- Docket | ROI Calculator: An interactive ROI calculator that helps prospective Docket customers quantify the pipeline impact of replacing legacy web forms with Docket’s AI conversational agent, using their own traffic volume, deal size, and close rate as inputs.
- Okta | Identity Threat Protection: A presentation designed for high-impact direct enablement for an audience of new account executives who have little to no experience with this real-time security product.
- Nutanix | Acropolis Hypervisor: An interactive module designed for self-guided exploration of an enterprise hypervisor solution created to eliminate the need for third-party hypervisors and to provide simplified management, disaster recovery, and automated security.
Analysis
These are modules I created for the MIT xPRO course (14.310x) called “Data Analysis for Social Scientists”.
- Module 01: Part 1: Bar Charts & Leaflet Maps
- Module 01: Part 2: Scatterplots
- Module 02: Fundamentals
- Module 03: Histograms, Kernel Density Estimates, & Joint Density Plots
- Module 04: Markov Inequality Simulation
- Module 05: Distributions and qnorm()
- Module 06: Maximum (Log) Likelihood Estimation
- Module 07: Nonparametric Regression
- Module 08: Linear Regression
- Module 09: Fuzzy Regression Discontinuity Design
- Module 10: Design Principles of Data Visualization
- Module 11: Random Forest Regression & Sentiment Analysis
- Module 12: Review: Sharp Regression Discontinuity & Nonparametric Regression
These instructional analyses were created for a teacher professional development program at the Learning Systems Institute at Florida State University. The goal was to provide concrete, practical examples of data analysis for science education to help teachers learn how to implement scientific analyses in the classroom using open-source programming tools that were also used by practicing scientists.
This Fountain Pen Ink Analysis represents a quantitative analysis of over 500 fountain pen inks using over 15,000 user-provided reviews from The Goulet Pen Company.
Didgeridata is a multidisciplinary outreach experience that was created for MIT to fuse music, engineering, data exploration, and storytelling.
Productivity
- [GitHub] In the course of my work at Emeritus.org and ZoomInfo, I created custom productivity tools for processing a variety of documents.
Creative Coding
These are some fun vibe-coded interactive and engaging visualizations. The Wall Art example was developed while exploring ideas for an article I co-authored for MIT’s Computational Law journal.