Logan P.

I design and build software from scratch, and I own it end to end: from the first spec to a system running in production. Some of what I have built from scratch: • A real-time decisioning engine that replaced a vendor platform: an HTTP service with a configurable vendor waterfall, a fully auditable decision log in PostgreSQL, and a replay harness for testing changes safely. It is now the core engine of a lending technology company. • A deterministic physics and simulation engine in C#/.NET, with lockstep multiplayer networking, a renderer and a level editor, using no third-party engine, physics or maths library. • A language model: architecture, tokenizer and training. • Production scoring pipelines (two models, calibrated and backtested) that feed live decisions every night. • Cloud services, integrations and large data migrations for a nonprofit's SMS health platform, including backfills of hundreds of thousands of records in BigQuery. • Data warehouses with nightly multi-source pipelines, and the dashboards on top of them. What I take on: new builds (back-ends, APIs, integrations, internal tools, data platforms, automation and AI agents), and systems that are stuck: databases that freeze or lose data, half-finished apps, code nobody understands anymore. How I work: I start with a short fixed-price scoping or diagnosis, so you know the plan and the price before committing to more. I use proprietary AI tooling of my own, which lets me ship quickly and prove the work is correct. Stack: Python, SQL (PostgreSQL, MySQL), C#/.NET, JavaScript, Docker, Google Cloud (Cloud Run, BigQuery), Databricks, Azure.