Louis Forster

Applied ML and systems work. Recent projects include graph-based retrieval for collaborative AI software and offline local-LLM tooling for Jetson devices.

Projects

  • Turns Excalidraw canvases into a queryable knowledge graph with semantic, relational, and spatial retrieval.

    • - Built a canvas-to-graph ingestion pipeline for Excalidraw data so retrieval can use structure instead of flat text
    • - Production deployment supporting concurrent collaborative sessions with 13M+ LLM tokens processed

    TypeScript / React / Excalidraw / Node.js / Graph database / RAG

  • Offline local-LLM terminal app for Jetson and edge Linux with on-device models, agent tools, and explicit safety gates.

    • - Published as the `open-jet` package for Jetson and edge Linux
    • - Runs fully offline after local model setup via llama.cpp

    Python / Textual / llama.cpp / GGUF quantization / LLMs / Nvidia Jetson

Experience

  • Visual Computing Research Intern

    Bristol Vision Institute

    Jun-Jul 2025

    Fine-tuned RAFT/DPFlow optical flow models using Sintel dataset with synthetic noise. Implemented bidirectional warping and forward-backward consistency checks for adaptive occlusion masking into the main denoising model. Stabilised the self-supervised training pipeline, achieving 29.22 dB PSNR (+0.13), 0.863 SSIM (+0.010), and 0.334 LPIPS (-9.2%) on BVI-RLV, surpassing prior state-of-the-art

  • Teaching Assistant

    University of Bristol

    Sep 2023-Apr 2025

    Teaching Assistant for up to 200 undergraduates across Linear Algebra, C, Discrete Mathematics, and Data Science

  • Software Engineer Intern

    Scribblepad Press

    Jul-Sep 2024

    Managed web deployment and site optimisation. Increased organic traffic by 119%.

Contact: louisforster64@gmail.com