Building software for Azure's datacenter supply chain: capacity planning, provisioning, and plan optimizations.

I joined Betteromics quite early (10th team member) to help them build their Lab Inventory Management System (LIMS).I worked alongside three other engineers to build out the first version of the LIMS. This was fullstack work and included designing some of the core concepts that the platform would be built around. For example, I designed the configurable attribute system that customers could use to define their lab's data model.As the product matured, I went on to build a system for bulk data-entry. This was an intelligent spreadsheet-style interface that included complex validation logic, computed fields, and a configurable appearance (i.e. showing/hiding columns in certain contexts). This was a challenging problem that involved a fair amount of performance-tuning to ensure that we could handle large amounts of data.As one of the few members of the team with a strong background in frontend development, I helped others ramp up on React and Typescript and built components and utilities that they could use to build new UIs quickly.During my time at Betteromics, I was surrounded by a group of extremely experienced engineers and leaders whose deep industry experience taught me a lot about planning and executing large software projects.

Clever connects users with real estate agents to help them save money when buying or selling a home.I hired and led the company's engineering team. (software, not construction 😉)When I joined, there was no technology infrastructure.There was a bit of "glue code" that one of the cofounders had written and was running out of a Jupyter notebook. The company database was a Google Sheet and a Hubspot account, and the team would comb through it to pick agents to match customers with.I built out a Django backend that we used to power our suite of internal tools. These helped match customers with agents and ensure that "in-flight" deals were checked up on regularly.I hired and led a team of 6 engineers.With few resources and a lot on our plates, we had to stay focused and make intelligent tooling choices to amplify our productivity.During my time here, we grew the sales team 5x and reduced the average "time to match" from 1.5 hours to just 12 minutes.I promoted a culture of sitting alongside the sales team and watching them work. Feeling and internalizing the pain-points of our stakeholders helped us prioritize our work for maximum impact.This was a fast-paced and immensely rewarding experience.

Role-based access controls, permissions, Javascript/React, Python/DjangoThis was a short freelance project where I helped Emburse refactor it's RBAC system to allow for more fine-grained controls as they scaled their platform's capabilities.

Gigster is a platform that matches top freelancers with enterprise companies.During my time here, the company went from 10 to 100 employees.My experience was characterized by "doing the dishes": volunteering to fix things that needed fixing.I was hired for my background in statistics and machine learning, and I ended up doing mostly fullstack web development because that's what the company needed most. I learned a lot in the process.----I was hired as one of Gigster HQ's first engineers. (I was a full-time employee at the company, not a freelancer on the platform.)Projects I'm particularly proud of:I updated the company's payments infrastructure, saving the company over $100k per year in transaction fees. This also streamlined the payments process and drastically reduced the time required to pay freelancers.I unblocked over $10M in sales pipeline as the interim security point-of-contact. This involved describing the company's security posture to prospective customers and proposing the technology team's security roadmap.I was a contributor to the internal platform that was used to manage the company's active projects with over 600 freelancers.I participated in an all-hands-on-deck project to change our primary database from MongoDB to Postgres.

Assisted in the development of a new offering for investment research organizations, processing satellite imagery to quantify industrial activity on both nationwide and regional levels.

Palantir is a data and technology consulting firm that was rapidly scaling at the time.My team’s client was one of the world’s largest financial services firms.I ran email marketing experiments on an email list of millions of small business owners to reduce churn in the client’s customer base.

In college, I cofounded an ad agency and scaled it to $80k/month in revenue.We did influencer marketing, securing online endorsements for brands.My cofounders and I were early into the space. We realized we could pay popular twitter accounts to post links to products.Because the space was so new, we were able to generate clicks for extremely cheap and outperform our customers’ other marketing channels. It was an arbitrage opportunity.My role was equal parts sales and ad-buying.SalesI conducted 200+ calls with potential customers, asking them about their marketing goals and strategies.This helped us understand our customers deeply and dial in our pricing for high margins.Ad-BuyingOnce customers agreed to work with us, we had to actually generate clicks on their links.I became really good at finding influencers in niche spaces and estimating how many clicks they would generate through sponsored posts.As we grew, I operationalized my process and managed an ad-buying team of four.