Developed and deployed AI-driven research tools that increased research speed by 2x-100x for different clients. Check out https://www.mokoresearch.com for more info and past projects. Some Highlights Include:- AI-powered Podcast Searcher: Created an AI tool automating podcast discovery, transcription, and content analysis, improving research speed and relevance.- Perspectives & Debate Generator: Helped build a website in Next.js that would take a question, generate perspectives on that question, and show arguments for each perspective.- Political Candidate Profiler: Built a tool to streamline the identification of local political candidates' online presence, including campaign website, email, and social media.- Scholarly Article Citation Crawler: Engineered a system to find the most cited papers in a field, along with the network of papers each of those papers cite.- Newsletter Sponsor Identifier: Automated the process of identifying sponsors in newsletters linking results directly to company profiles.- Political Forecasting Bot: In spare time, built a bot that can forecast future geo-political events. As of the last update, it won $900 in the Metaculus Q3 AI Forecasting Bot Competition. See https://mokoresearch.streamlit.app/. Later did a project with Metaculus to create a Base rate researcher and Key factors researcher.

During college worked to create the "Wikipedia for debatable topics". Originally organized college political clubs and high school debate students to crowdsource content into a custom site. When AI exploded, created software that scans the web for scholarly arguments for and against a given topic. It summarized arguments from both sides and cited direct quotes (see https://logipedia.co). Started bootstrapping Logipedia full time after college, then later contracted for 'The Society Library' who is working on a similar goal/mission and already had funding. The following was accomplished (while also finishing school):- Reduced manual research time by 50%-75% depending on the article - Content sent via newsletter to a group of 229 people and got open and click-through rates 3x industry normal (Though failed to find a good growth strategy as people were not sharing)- Conducted 144 market research interviews with a target audience to inform problem scope, software design, and product iteration- Organized 5 volunteers to help code the automation software

Created a 21-page document for WWT's sales team documenting the Agile Transformation Team's services. This document helped sales sell the team's services more efficientlyLed a team of interns to design a better feedback system for WWT and presented the solution to executives in a design competition.Facilitated 13 group brainstorming meetings and many other smaller meetings between both the Agile Transformation team and the intern teamWhen brainstorming sessions were slow, invented a meeting facilitation tool in Coda that allowed the team to get the needed information in half the time.Proactively kept the sales documentation project going and on time despite the main manager needing to leave and there being confusion about the direction of the project in his absence

Gave important insights about how to optimize memory bandwidth and latency on JMA’s Linux servers.Determined benefits and detriments of different OS, BIOS, and tuning combinations to maximize memory performance. This informed important company decisions about how to update JMA’s server configuration.Researched and determined multiple key benchmarks to measure memory performance. Created a large script to run the benchmarks and compile data into an easy-to-read CSV for analysis by company executives

Helped Ascension smoothly transition email between Microsoft Tenants by creating an email transfer solution using scripting and a written guide.Set up Microsoft Teams and user profiles on the company's new Microsoft Tenant. Parsed log files into a readable format using PowerShell (PowerShell was self-learned during the internship)Researched and gave a presentation on JFrog Artifactory to the company. This allowed for a deeper understanding of the “what, why, and how” of Artifactory before the company adopted it.