Hi, I'm Kashish. I'm a Cornelius Vanderbilt Scholar studying computer science and applied math at Vanderbilt, and I build AI agents.
Right now I'm interning at ID.me, where I built Member Assistant, a conversational AI agent that answers member questions across three of their products. It runs on Google's Agent Development Kit as a single Spring Boot service, and grounds its answers in over 10,000 offers across 500+ stores using Vertex AI Search, Elasticsearch, and Postgres, reaching 94% answer relevance. Before that I rebuilt the backend for a learning platform used by 3,800 students at Be The Source, and built a RAG pipeline at Trigyn.
Outside of work I've built a few things I care about, like FinCoach AI, a bilingual app that teaches students about money, and a financial literacy curriculum that got taught in schools across New Jersey. I'm also broadly interested in machine learning and where it's headed.
I care a lot about philosophy too, especially ethics and questions about how we ought to live, and it shapes how I think about the technology I build. When I'm not coding, you can find me bouldering, hiking, shooting film photography, or singing.
I built Member Assistant, a conversational AI agent that answers member questions across three ID.me products from a single Spring Boot service on Google's Agent Development Kit. I designed a tool layer that lets it pull from Vertex AI Search, Elasticsearch, and Postgres, grounding answers in over 10,000 offers across 500+ stores at 94% relevance, and cut p95 latency by about 35%. I also shipped the first store search and offer retrieval in the shop API.
I rebuilt the backend for a learning platform used by 3,800 students. I wrote 15+ REST endpoints, moved the data onto Postgres, which saved the org about $10K a year, and built a live attendance tracker on top of the Zoom API.
I built a RAG pipeline over 100+ internal documents, from ingesting and embedding the docs to serving answers through a production API. It cut support tickets by about 30% and answered correctly around 92% of the time.
The AI agent I built at ID.me. It decides which tool to call, pulls real offers and cashback rates from Vertex AI Search, Elasticsearch, and Postgres, and answers from that data instead of guessing. It covers over 10,000 offers across 500+ stores at 94% answer relevance.
Explore the interactive breakdown →A bilingual app I built that teaches students about money. It walks 50+ students through budgeting, banking, and getting an ID, and I'm turning it into a curriculum a few New Jersey high schools can use.
A tool for checking how well a RAG or agent setup answers questions. It runs a fixed set of questions against a synthetic document set, has an LLM grade the answers, and checks those grades against my own labels so the scores mean something. It runs fully on local inference.
A tool that pulls all my deadlines from Gmail, Calendar, Slack, and Canvas into one place, so I stop missing things across school and work.