Julius M. Schmidt

rooted (formerly ArtSlide) | JJLab UG

Co-Founder

Co-founded a sustainability-focused venture as a two-person founding team, building digital marketplaces for locally made home goods. Responsible for finance, strategy, product and analytics, alongside my studies.

Financial modelling

Built the company's financial models: the plans behind two friends-and-family rounds and a driver-based monthly model (from traffic and conversion to income statement, CAPEX and working capital).

Strategy & OKRs

Wrote the company strategy and the marketing strategy (segmentation, targeting, positioning); ran quarterly OKRs.

Product & analytics

Owned the product roadmap, pricing tiers and A/B tests; product analytics in Mixpanel and marketplace KPIs such as supply and demand liquidity.

Processes & finance operations

Modelled key processes in BPMN (e.g. payment flow); prepared every annual statement and the bookkeeping in-house.

Customer discovery & go-to-market

Interviews with customers and makers, competitor analysis and launch strategies for both sides of the marketplace; steered freelance customer-success support (2023 – 2024).

vertigo

Personal AI system

A self-running AI monitor for my own investments. Every morning a GitHub Actions job checks each holding against the thesis I wrote for it, and every line it produces is marked as calculated or written by a language model.

job-triage

Personal AI system

A daily pipeline for my own job search. It scores postings against a written profile with Claude, puts the few worth a look on a Notion board and learns from my decisions; a stop-list runs in shadow mode before it acts.

Bachelor thesis

Benchmark study

Do liquid neural networks forecast daily financial returns better than classical models? 11 variants against 8 baselines on ten years of AAPL, JPM and SPY data, walk-forward over 5 seeds, Diebold–Mariano tests with Holm correction. Result: no, none of the 120 comparisons was significant. Code, method and thesis are public.