I’m Jenil Shah, a Staff Machine Learning Manager based in Seattle. My work sits at the intersection of recommendation systems, personalization, and AI. I’m interested in how we turn ideas into useful software, and what we learn when those ideas meet the real world.
I tend to learn by doing: trying a tool, writing a small piece of code, or following a question further than I intended. Writing helps me work through what I’ve found.
Work
At Amazon, I’ve worked on book recommendations and personalization, including the Multi Entity Similarities team, personalized reading experiences, and the data platforms behind them.
Earlier, I worked on dynamic ranking, skill arbitration, and domain classification in Alexa AI. Before that, at Ford, I built simulations for vehicle features using Qt, MATLAB, and Simulink.
Outside work
I tinker with machine learning, LLMs, productivity tools, and financial data. I like finding ways to make everyday tasks a little less cumbersome. Away from the screen, I enjoy getting outdoors.
Why I write
This site is a record of what I’m thinking about and learning. My writing covers engineering, AI, productivity, and the occasional idea that doesn’t fit neatly into a category. Some experiments get their own project space.
Docket is my project for self-hosted AI decision models: classification, routing, scoring, and agent orchestration. Its project pages include benchmarks and limitations and an API reference.
Community
I’ve served on program committees for RecSys, IISA, and IEEE BigDataService, and judged an Open Source AI Hackathon at Microsoft Reactor in Seattle.
Get in touch
Email me at [email protected], or find me on LinkedIn.
The opinions on this site are my own and don’t represent my employer.