Joel Williams

I am currently studying math, building software, and training Judo.
I enjoy making games as a hobby and I'm interested in economics, statistics, and predictive models.

Projects

[RASCII]

Converting images into ASCII art in Rust

rascii 1
[1/2]
rascii 2
[2/2]

Product bullets

  • Implemented a fast, memory-efficient image→ASCII renderer in Rust that preserves visual detail using aspect-aware resizing, luminance mapping, and error-diffusion dithering.
  • Streamed output and parallelized per-row processing to minimize memory overhead and accelerate transforms on multi-core machines.
  • Supported optional ANSI-colored output by mapping image blocks to terminal colors while keeping the pipeline O(n) in pixels.

Technical highlight: Using bilinear resizing plus Floyd–Steinberg dithering and per-row parallelism yields high-quality ASCII output with linear runtime and low memory usage.

STACK: RUST

[NEMO]

A people and company research tool for sales teams and recruiters

rascii 1

Product bullets

  • Built a modular, tools-based Python backend that orchestrates multi-source web search and content extraction to power an agentic company-research application.
  • Designed a pluggable Tool interface so new data sources (YouTube, LinkedIn, filings, etc.) can be added with minimal changes to orchestration.
  • Implemented parallelized fetch-and-merge workflows and chunked text processing to produce timely, summarized results for frontend consumption.

Technical highlight: The architecture centers on a lightweight Tool abstraction and async orchestration that minimizes latency and lets the system scale by adding new Tools without touching core orchestration logic.

STACK: NEXT.JS, PYTHON, and TYPESCRIPT