Ecommerce · 2026
LLM features with RAG
Integrated LLM-powered features into a production ecommerce product, using retrieval-augmented generation to ground answers in real merchant data with citations.
Available for AI/LLM projects · 2026
Independent senior engineer in Austin. I help SaaS founders integrate LLMs, RAG, and AI features into production — direct work, no agency overhead. Currently building Pipelane, an AI marketing platform for home-services businesses.
Selected work
Ecommerce · 2026
Integrated LLM-powered features into a production ecommerce product, using retrieval-augmented generation to ground answers in real merchant data with citations.
Marketplace · 2026
Co-built end-to-end — a live marketplace connecting students with advisors who've walked the same education or career path. thinkupsocial.com
Creative tools · 2025
Shipped real-time search over multimedia documents. Lifted user retention 20% on the product it landed in. Foundational pattern for retrieval pipelines.
Streaming · 2022
Built a Netflix-style recommendation engine — collaborative filtering that tailors content to each user based on their activity in the app.
Streaming · 2024
Shipped features across frontend, backend, and infra that helped grow the product to 60,000+ customers in its first year.
Creative tools · 2024
Led a team of three engineers to build and integrate a resizing service that significantly cut latency across the product's digital assets.
Start a project
The best fit is usually a 30-minute call where you describe the problem and I ask annoying questions. There's no pressure — if I'm not the right person, I'll say so and point you somewhere better.
FAQ
AI features that actually have to run in production, not just demo well. LLM integration, RAG, evals, agents — usually a single well-scoped feature in a 2-4 week sprint. If you want me to "rewrite your whole product with AI," I'm probably not the right fit.
Totally fine. The first conversation is usually me asking annoying questions until we both understand what you actually want and what it'll cost. You don't need to know what an embedding is.
Direct. You talk to me, not a project manager — I'm the one writing the code. I prefer fixed-scope sprints over open-ended retainers; keeps both of us honest about what's actually getting built.
Email vincentvais17@gmail.com or use the form above. I usually reply within a business day.