Hi, I'm Will Boudy, and I'm currently working on a Master's in Computer Science at Johns Hopkins University.

Focused on orchestration, deep learning, and distributed systems.

Here's some of my work.

Will Boudy

Work

  • >
    Dynamic Titles at Amazon

    Built a LLM-driven title generation system and shipped experiments to 10K+ users.

    Experimentation
    Catalog Personalization
    Summer 2025

Research

  • >
    Forecasting with Dr. Andrews

    Training deep research forecasting agents for general event prediction.

    Probabilistic Forecasting
    Benchmarking
    Ongoinglab
  • >
    Biopharma with Dr. Betenbaugh

    Genetically engineered CHO cell lines for enhanced monoclonal antibody (mAbs) production. Mentored a student through a summer project on fungal bioleaching of metals.

    CHO Engineering
    mAb Production
    2022-2024lab

(Pinned) Projects

  • >
    Mycelium

    Personal knowledge vault and Obsidian-style workspace for research and development notes.

    Note Systems
    Research Logs
    Ongoinggit
  • >
    Tally

    Voice-first logging assistant that turns spoken updates into spreadsheet-friendly entries and lightweight tracking.

    Speech Input
    Tabular Tracking
    Ongoing
  • >
    Fly-prompts

    Curated prompts for orchestrating multi-agent coding workflows with Emanuel's Flywheel.

    CLI Interop
    Execution Patterns
    One-hour projectgit
  • >
    APR Control Plane

    Wrapper around automated_plan_reviser_pro for coordinating multi-agent workflows across environments.

    Supervisor Layer
    Remote Execution
    One-day projectgit
  • >
    Entropy Wiki

    Reference hub covering coding agent environments, orchestration frameworks, and reusable AI skills and prompts.

    Agent Ecosystems
    Skill Library
    One-week projectgit
  • >
    Antigravity Crew

    Human-gated agent workflow tailored for the Google Antigravity IDE, with clear planning, implementation, and verification handoffs.

    Human-Gated Flow
    Auditability
    One-week projectgitnotes
  • >
    Research Intelligence Engine

    RAG assistant featuring a two-stage retrieval pipeline: semantic vector search followed by cross-encoder reranking to provide cited answers across 100K+ arXiv papers.

    Scholarly Search
    Evidence Synthesis
    Two-week projectgit
more here

Open Source

  • >
    Clawdbot Cost Monitor
    • #4: Improved security by making the dashboard localhost-only by default.
      merged
  • >
    Codex
    • #11052: Proposed adding `skill_name` so events can be attributed to the active skill for ROI tracking.
      Issue
  • >
    Openclaw
    • #14308: Passed `docker.env` into sandbox container creation for predictable runtime config.
      open
    • #15859: Improved graceful fallback and model-failure logs for faster diagnosis.
      open