This isn’t an exhaustive catalog, but rather a curated set of projects that showcase the breadth of challenges I’ve taken on and the creativity I bring to solving them.
What you’ll find: a sampler of work across industries and product types—0-1 marketplaces, AI foundation models, sunsetting legacy products, new launches for multiple user types, research artifacts, cross-company initiatives, AI-integrated side projects, and yes… even bananas. 🍌
AI, EdTech, Marketplaces, Personal Projects & Bananas
Generative Consumer AI Model
Generative Consumer Model MVP
As Product Manager at a new MIT startup, I led the development of the first demo of our foundation AI model—the Generative Consumer Model (GCM). Built in just one month with a lean team, the MVP showcased how probabilistic AI could impute missing data, generate probabilities, and surface insights that traditional models could not. We delivered a live demo for field presentations that made the power of probabilistic modeling tangible for partners and customers.
-
Company: SaferData
-
Product: Generative Consumer Model (AI)
-
Time: 1 Month
-
Phases: 2
-
Team: 2 Engineers, 1 Data Scientist, 1 Contractor, 1 Product Manager (self)
.jpg)
Learning Tool Interoperability (LTI) Configuration
A cornerstone initiative designed to strengthen platform stickiness by enabling seamless integration of third-party learning tools. With providers and institutions both relying on these integrations, the project had to balance two priorities: making it simple for providers to add and manage their configurations, and ensuring end users could easily configure and deploy them in their learning environments.
-
Company: Instructure
-
Product: Canvas, SysAdmin, Partner Portal, Marketplace.
-
Time: 7 months
-
Phases: 3
-
Team: 10 engineers, 2 designers, 2 contractors, 1 PM (self)
_edited.jpg)
Partner Portal
I led the full lifecycle of the Partner Portal launch—from research and development through GTM execution. Designed as the single source of truth for third-party partners, the portal enabled centralized asset management across five interconnected products. The initiative required solving for cross-product synchronization, introducing new revenue-generating partner features, and orchestrating a simultaneous launch alongside our new Marketplace website while deprecating legacy portals. Delivering this successfully demanded phased planning, and coordination across many products' teams.
-
Company: Instructure
-
Product: Partner Portal
-
Time: 9 months
-
Phases: 4
-
Team: 5 Engineers, 1 Designer, 1 Product Manager (self)
Custom Created Products
This new feature launch redesigned the product request workflow by meticulously mapping user flows to rethink a feature launch that reworked both customer and internal processes. The solution eliminated major pain points, enabled phased rollout without breaking systems, and increased product stickiness by giving customers more control, transparency, and speed.
-
Company: LearnPlatform
-
Product: LearnPlatform
-
Time: 6 months
-
Phases: 3
-
Team: 4 Engineers, 3 Contractors, 2 Designers, 1 Product Manager (self)
Product Market Fit...with no product
How do you perform research with just a technology, and isn’t yet a product? I designed a modular discovery prototype to explore the core value of a Generative Consumer Model. By combining market analysis with user research, the project validated product, market, and customer assumptions—helping identify viable paths to product–market fit
-
Company: SaferData
-
Product: Generative Consumer Model (research)
-
Time: 1 month
-
Phases: n/a
-
Team: (self)
TwinTakes.ai - Dual-Perspective AI for Product Managers
I independently designed, built, and deployed TwinTakes.ai, as a personal passion project. The tool provides two simultaneous AI models to aid in decision-making:
Builder LLM: a growth-oriented voice that advocates for expansion and feature innovation.
Inverter LLM: a contrarian perspective that challenges assumptions and encourages simplification.
This dual-lens approach helps product managers explore trade-offs in areas such as feature prioritization, pricing, user experience, and growth vs. retention strategy.
-
Company: Personal Project
-
Product: TwinTakes.ai
-
Time: 1 week
-
Phases: n/a
-
Team: Self
Bananify
Bananify.info is an app that translates ordinary measurements—weight, distance, even time—into the only unit that really matters: bananas. Built start-to-finish with a terminal based AI (Warp). For me, it’s both a practice in nurturing creativity and a way to stay sharp with the newest AI building tools shaping the tech landscape. This was a great personal project to expose me to the technology before progressing to more rigorous application builds.
-
Company: Personal project
-
Product: Bananify
-
Phase: n/a
-
Team: self
My Interactions
My Interactions is a personal project! This web app helped users privately check for potential drug-to-drug/supplement interactions. After creating a secure account, users could enter two substances, view possible interactions powered by RxNorm and Drug Interaction APIs, and save past searches for future reference. The app was built as a hobby project using Node/Express, MySQL, Sequelize, and APIs from the National Library of Medicine, and deployed on Heroku
-
Company: Personal Project
-
Product: My Interactions
-
Time: 1 month
-
Phases: n/a
-
Team: Self
*The original API used (RxNav) was free, but per the National Library of Medicine’s notice, RxNav discontinued its drug–drug interaction features on January 2, 2024.