CSE student building ML systems and the backends that keep them running.
I'm a second-year Computer Science student. I'm interested in the space between machine learning and backend engineering: taking a model that works in a notebook and turning it into a service people can actually use.
I started with C++ and data structures, and these days I'm working with Python for ML and Java for backend services.
Last updated: November 2026
A workspace-based RAG app: upload PDFs and notes, then ask questions and get answers that cite the exact passages they came from. Documents are chunked, embedded with a local model and searched with pgvector, and a local LLM writes the answers, so nothing leaves the machine.
A job runner for ML workloads. You submit evaluation or batch-inference jobs over a REST API, workers claim them from a Postgres-backed queue, failed jobs retry with backoff, and progress streams live to the browser. The first job it runs will score Second Brain's answers against a fixed question set, so I can catch regressions when I change the retrieval.