I build applied-AI products that turn messy source material into grounded decisions and usable outputs. I am looking for contract, fractional, full-time, or credible co-founder opportunities, especially research products, document intelligence, and internal decision workflows.
Recent work includes a Colorado-only funding decision-support MVP: public program sources become normalized records, ranked recommendations, and reviewable Markdown/PDF briefs. The recorded evidence includes 132 records, 750 extracts, a 30-case quality review, and two end-to-end pilot workflows. These are MVP/pilot results, not claims of production adoption.
I also built an invite-only research companion with source-linked interaction and private workspaces, and helped build Saffron Threads (https://saffron-threads.com), with AI-generated fashion models and a client-facing backend for generating and managing product imagery.
For contract work, a useful first engagement is one bounded workflow with clear acceptance criteria, followed by a scoped prototype or build. Happy to discuss the source material, user needs, and intended output through LinkedIn.
Remote: Yes
Willing to relocate: Yes
Technologies: Python, FastAPI, PostgreSQL, SQLAlchemy, Docker, Playwright, structured AI output, data extraction and normalization, embeddings, ranked matching, authenticated web applications, Markdown/PDF reports
Résumé/portfolio: https://github.com/TashaSkyUp
Contact: https://www.linkedin.com/in/tasha-upchurch
I build applied-AI products that turn messy source material into grounded decisions and usable outputs. I am looking for contract, fractional, full-time, or credible co-founder opportunities, especially research products, document intelligence, and internal decision workflows.
Recent work includes a Colorado-only funding decision-support MVP: public program sources become normalized records, ranked recommendations, and reviewable Markdown/PDF briefs. The recorded evidence includes 132 records, 750 extracts, a 30-case quality review, and two end-to-end pilot workflows. These are MVP/pilot results, not claims of production adoption.
I also built an invite-only research companion with source-linked interaction and private workspaces, and helped build Saffron Threads (https://saffron-threads.com), with AI-generated fashion models and a client-facing backend for generating and managing product imagery.
For contract work, a useful first engagement is one bounded workflow with clear acceptance criteria, followed by a scoped prototype or build. Happy to discuss the source material, user needs, and intended output through LinkedIn.