Skip to content
View Cyriellewu's full-sized avatar

Block or report Cyriellewu

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Cyriellewu/README.md

Hi, I'm Willow 👋

M.S. Computer Science (Machine Learning) at Columbia, graduating December 2026. I work on NLP, knowledge graphs and ML interpretability, and I build the data pipelines and tools around them.

Research

  • Explainable Dynamic GNNs for Predictive Maintenance in Vehicle Chassis Systems. First author, ACML 2025 (PMLR 304). paper · repo
  • FKGE-OSC: Fuzzy-Enhanced Uncertain Knowledge Graph Embeddings. First author, manuscript. repo
  • Research Data Analyst, Columbia Business School (Summer 2026): entity resolution over billion-row finance datasets

Selected projects

  • llm-lab: local-first, DuckDB-native workbench for evaluating LLMs and agents
  • duckdb-torch-pipeline: out-of-core Parquet → DuckDB → PyTorch streaming DataLoader
  • evidenceoflife-v2: visual day planner with draggable timelines and map-based memories

Open source

Stack: Python · PyTorch · PyG · DuckDB · TypeScript / React

📫 yw4643@columbia.edu · LinkedIn

Pinned Loading

  1. evidenceoflife-v2 evidenceoflife-v2 Public

    ADHD-friendly visual planning with draggable timelines, planned-versus-actual tracking, and map-based memories.

    TypeScript 2 4

  2. duckdb-torch-pipeline duckdb-torch-pipeline Public

    Out-of-core Parquet->DuckDB->PyTorch streaming DataLoader with pushdown, Arrow-native tensors, bounded-memory shuffle, and a benchmark vs pandas full-load

    Python

  3. llm-lab llm-lab Public

    A local-first, DuckDB-native workbench for evaluating LLMs and agents: suites, multi-model runs, pluggable scorers (incl. SQL-execution + drift), and a dashboard.

    Python

  4. oracle-os oracle-os Public

    A React/TypeScript reflection app with deterministic symbolic engines, a local journal, and optional LLM interpretation.

    TypeScript

  5. explainable-gnn-predictive-maintenance explainable-gnn-predictive-maintenance Public

    ACML 2025 (PMLR 304) · Explainable hybrid GCN + Random Forest for predicting high-cost vehicle faults from SPN/FMI fault-code graphs. 98.9% accuracy.

  6. fuzzy-uncertain-kg-embeddings fuzzy-uncertain-kg-embeddings Public

    FKGE-OSC: uncertain knowledge graph embeddings with fuzzy logic, a Lee oscillator layer and PSL for confidence-score prediction.