We’re building an AI fashion styling agent: a revolutionary assistant that curates your wardrobe, plans incredible outfits, and turns shopping into a confident, effortless experience. This is all made possible by cutting-edge deep learning vision models and agentic AI. Stay tuned!

We build software that is a launchpad for growth. It solves customer problems and adapts to your evolving business. We work hand in hand with you and your customers. We use agile practices and commit to technical excellence.We specialize in applied AI and LLM-based app development. We can jumpstart your AI strategy by building your LLM app. You won't need to hire a full-time software team.If you have a software team, you can avoid a big commitment on a tech leader. Bring us on as your Fractional CTO. We also offer standalone advising. It will help you get the most from your software development team.

In this role, I incubate and develop AI-powered startups, each targeting specific industry challenges with advanced technologies, including large language models and retrieval-augmented generation (RAG):HomeHub – Created an AI solution that navigates federal, state, and utility incentives for energy efficiency, assessing consumer eligibility, stacking incentives, and providing cost estimates based on available rebates and incentives. HomeHub leverages commercial LLMs and integrates multiple datasets, including EIA geospatial utility data, HUD income data, and U.S. Census data, to simplify incentive discovery and maximize savings for consumers.OneGC – Developed an AI-powered legal platform tailored to startup founders, automating 90% of routine legal work. OneGC enables founders to ask frequent legal questions without incurring high legal fees. Utilizing large language models, RAG, and integrations with corporate systems, the platform delivers on-demand legal assistance at a lower cost.HelloBonjour – Built an chat AI interface that connects various corporate data sources, allowing employees to query information about the company across multiple systems in real time. HelloBonjour integrates commercial large language models with RAG, enabling seamless access to information from diverse corporate systems through a single chat interface.

Led and grew the Software Engineering, Data & Data Science, and Product Management teams within Maisonette. We built customer marketplace technology that made it easy for busy parents to find the best products for their children.We built a data warehouse from scratch that serviced the companies Business Intelligence needs. We also applied data science techniques to derive insight from Maisonette shopping and customer behavior data.

I lead software engineering, IT, and security at Casper. Our 75 person Technology team is mostly located in NYC, but we have team members in our Berlin, LA, and SF offices. We build a lot of the software that powers the Casper business: our open source Ruby on Rails e-commerce platform, a multi-content fulfillment and logistics network, integrations with our retail partners (Amazon, Target, etc.), the point of sale iOS app that we use in our retail stores, the content management system behind casper.com, a data warehouse with a visual interface for self-service data analysis, and the CRM that is the backbone to our customer service, marketing acquisition, and retail lead management.We have a service oriented architecture hosted on AWS. Most of our backend applications are Ruby with a bit of Python and Elixir. On the frontend we make heavy use of React and Redux for our server rendered, static frontends.

I started as the first engineer in April 2015 and helped build a Tech team of about 80 in a year-and-a-half. Our team built the fundraising frontend for hundreds of millions of dollar in donation and millions of transactions. We built digital products to help millions of volunteers across the country. We architected and built the content management system behind hillaryclinton.com that had a near real-time TTL for content changes while also scaling traffic. We managed a large Vertica database with data for millions of constituents. We built an amazing mobile app that kept voters and volunteers engaged with the campaign. And we did so much more!We built all of this in a micro-service architecture environment hosted on AWS. We had over 90 services in production. We primarily used Python, Ruby, React, NuclearJS. We took advantage of Lambdas to simplify the backend and keep most of our frontends static.

I lead optimization efforts throughout the company including the marketing website and product. We used Optimizely to improve Optimizely. My small marketing engineering team refactored optimizely.com making it static and hosted on Google App Engine. We ran experiments that improved our return on ad spending by $1 million a year. We integrated Optimizely with several partners like Heap and Salesforce (B2B sales workflow) to drive a deeper understanding of the results of our experiments.

Managed the online fundraising platforms that brought in $690 million during the course of the campaign. This included architecting and building a new donation platform for the Digital team that raised $250 million in 6 months. Also worked as the product and technical lead for Quick Donate (one-click donations) which brought in $115 million over its lifetime. Our web performance efforts reduced page load time by 60% which resulted in a 14% increase in conversion rate. We ran over 500 a/b tests that improved the overall donation rate by 49% and the email acquisition rate by 161%.

Digital strategy and web development for clients ranging from the Human Rights Campaign, Freedom to Marry, Communication Workers of America, People for Bikes, and the 2009 Presidential Inaugural Committee.