AI Engineer
Computer Engineering graduate from NYU building AI agents, retrieval systems, developer tools, and open-source software.
About
I like building software where product taste and systems thinking meet.
I am a Computer Engineering graduate from NYU focused on AI systems, retrieval workflows, and developer tools that make complex software easier to reason about.
My work tends to sit close to the engineering surface: model orchestration, data flow, observability, interfaces, and the infrastructure needed to make AI products reliable.
I care about clean abstractions, fast iteration loops, and interfaces that help engineers inspect what their systems are doing.
Building
Pelmen
Learning
Mojo
Interested In
Inference
Exploring
Compilers
Using
Developer Tools
Skills
A compact stack across product engineering, AI infrastructure, systems, and deployment.
Selected Work
Three representative projects across agents, retrieval, inference tooling, and production engineering.
A TypeScript coding agent CLI built with Bun, OpenRouter, Commander, and Ink/React, featuring streaming tool calls, surgical edits, session restore, model switching, context tracking, summarization, and reusable markdown skills.
A local-first hybrid code retrieval engine that indexes repositories into SQLite with AST symbols, semantic chunks, embeddings, file metadata, incremental reindexing, exact/fuzzy/semantic search, MCP-compatible output, Streamlit UI, and Mojo-accelerated ranking kernels.
A heterogeneous Graph Neural Network built with PyTorch Geometric to identify devices in mesh networks using custom Device, Event, and Service nodes, hyperparameter sweeping, and weighted loss functions to reach 90% accuracy.