Tech
I’m an MS Computer Science student at Northeastern University focused on data engineering. These are the projects I’m building toward data engineering and data analyst internships for Summer 2027.
A metrics-first product analytics pipeline: Python ingestion, dbt models from staging to marts, Airflow scheduling, a warehouse star schema, data tests, and a dashboard on top.
My solo Python data-engineering layer on my fork of OmniRAG: corpus-scale incremental ingestion, Postgres + pgvector storage, data quality and lineage, and evaluation analytics.
A chain-query join-order optimizer that uses the same Θ(n³) dynamic program as matrix-chain multiplication, compared against greedy and left-to-right heuristics on four real SQLite databases (Northwind, Chinook, Sakila, F1DB). Every row count and block size is checked against the actual SQL joins. Built with teammate Yanglin “Elliot” Hu.
150.88× less intermediate work than left-to-right join order on the Sakila database
A domain-aware Retrieval-Augmented Generation (RAG) engine built with Java 21 and Spring Boot with teammates Rui Song and Siyuan Liang. Upload documents, ask questions in natural language, and get grounded answers with source citations.
My solo, from-scratch compiler in C for Jive, a small teaching language, targeting x86-64 NASM assembly. Built stage by stage: lexer, parser, stack-machine IR, symbol table, function calls, control flow, then strings, heap allocation, and arrays.
Coursework, experiments, and everything in progress live at github.com/johnha912.