Selected Work
Davids Archive
A personal reading archive I use every morning. 62 sources are chunked into a 1,021-piece corpus I can query in natural language, with generated quizzes and recall sessions so I actually retain what I saved.

Startup Idea Validator
Type one sentence describing an idea and get a competitive landscape back in under 90 seconds, built from live search across Tavily, Product Hunt, and YC rather than from whatever the model remembers. It ends on one ranked wedge with the evidence behind it.

Lokin
A platform for student founders to find collaborators, co-founded at UW-Madison and grown to 500+ signups. Only 5% of signups were finishing a profile, so I rebuilt onboarding from six steps to three and completion went to 50%.
I enjoy building AI systems for complex environments where the context is human and messy. I broker FX options at TP ICAP and have built AI/ML tooling that seven brokers on the desk use daily. This has meant translating front-office workflow problems into working systems for liquidity discovery and broker decision support, on flow carrying $1.5K to $2K a week in attributed commission. My other contributions have included order-routing automation and contract-pricing improvements.
I graduated from the University of Wisconsin–Madison with a degree in data science and an economics minor. While I was a student, I co-founded Lokin, a digital hub for student builders we took from zero to 500+ signups in the first months after launch. I am most interested in solutions engineering, technical GTM, and deployment work that combines customer context, commercial judgment, and hands-on building.
Outside of work, I've built an AI reading archive for the problem of reading a lot of digital content and retaining only a fraction of it. The system I have built helps address knowledge comprehension and digital curation. Everything I upload is chunked, embedded, and searchable, and the retrieval feeds a learn engine that generates quizzes and open recall questions graded against the source, chat that cites the pieces it answers from, and a spaced queue that resurfaces relevant knowledge before it fades. And my startup idea-validation system helps me quickly test and evaluate various product ideas, allowing me to filter out weak concepts before investing time.
Experience
Selected for a desk-embedded AI deployment mandate to translate front-office workflow problems into production AI/ML systems spanning pricing, order routing, liquidity discovery, and real-time broker decision support.
Provide real-time price discovery and market intelligence on G10 FX options while building data products and internal tools for the desk.
Rotated across five desks spanning equities, fixed income, rates, and FX before returning to the FX options desk full time.
Built the product from zero on a founding team of four, covering user research, roadmap decisions, and the full technical stack.
Analyzed post-trade execution costs for institutional equity portfolios in Python and SQL, isolating slippage and market impact.