lilguy.win is a fast, serverless internship search engine and aggregator indexing 8,300+ open roles across 1,000+ employers in tech, finance, consulting, healthcare, engineering, logistics, and more.
Explore the live feed at lilguy.win (mirrored at lilguy.pages.dev).
- ⚡ Sub-Millisecond Edge Architecture: Zero-backend static distribution powered by Cloudflare Pages. Indexes 8,300+ postings in an ultra-lean ~8 MB bundle with on-demand description shards.
- 🔍 Instant Search & Multi-Axis Filtering: Search role titles and company names with realtime debounce, combobox auto-suggest, and filters across Industry, Job Function, Workplace Type (Remote, Hybrid, On-site), Location, Recruiting Cycle, and Posting Age.
- 🏢 Deep ATS & Harvester Coverage: Ingests directly from Greenhouse, Lever, Workday, Oracle Recruiting Cloud, SmartRecruiters, Ashby, Schema.org JSON-LD microdata, and major aggregators.
- 📱 Responsive & Accessible UI: Mobile-first layout with native bottom drawer filters, auto-zoom prevention on mobile inputs, persistent pagination, full keyboard navigation, and WCAG AA contrast.
- 🤖 Application Autofill Suite: Browser userscript (
autofill/) that fills repetitive application fields from a local profile without auto-submitting. - 🎯 Forkable Filter Presets: Ready-made filter configs in
presets/for Software Engineering, Data & Analytics, Finance, Marketing, Sales, HR, and Supply Chain & Operations. - 📡 Filtered Feeds, Badges & Trends: Subscribe to any filter combo as an Atom feed, embed a live posting-count badge, or pull weekly hiring-pace stats — see Feeds, Badges & Trends below. No account, no email, no data collected about who's watching.
Everything here runs off the same static bundle everything else does — no signup, no API key, nothing stored server-side about who's asking.
Filtered Atom feeds — subscribe in any feed reader, or point an RSS-to-Discord/Slack/IFTTT bot at it for a poll-based alert with zero infrastructure on either side:
https://lilguy.win/feed.atom?category=Software+%26+Technology
https://lilguy.win/feed.atom?company=openai&job_function=Data+Science
https://lilguy.win/feed.atom?q=robotics&max_age_days=7&limit=50
Supported params: company (a key from /data/companies.json), category, job_function, work_arrangement, cycle_season, cycle_year, q (free text), max_age_days, limit (≤500). No params returns the default unfiltered feed.
Embeddable badges — a company's own open-role count, or the site total, as an SVG:
Hiring-pace trends — /data/trends.json publishes a 12-week weekly opened/closed series and the week's fastest-hiring employers, computed straight from postings' own first_seen/closed_at history (see service/edge_export.py's build_trends). The homepage footer surfaces the latest week inline.
Contextual share previews — a link to ?posting=<id>, ?company=<key>, or ?category=<name> shows that specific role, employer, or category's title/description and a matching share-image when pasted into Slack, Discord, or iMessage, instead of the generic homepage card (functions/_middleware.js + functions/og-card.svg.js, both Cloudflare Pages Functions). The per-entity image is SVG, rendered at the edge from the same static data the site itself reads — Twitter/X's card validator is pickier about non-raster og:image and may show no image there, a deliberate tradeoff against pulling a browser-rendering toolchain into this project just for image generation. The generic homepage share image (og-image.png) is a real PNG, regenerated on every export with that export's live open-role/employer counts (service/og_image.py, Pillow) — it used to be a hand-rendered screenshot with the numbers typed in as literal text.
┌─────────────────────────────┐
│ Continuous Ingestion / │
│ Automated ATS Discovery │
└──────────────┬──────────────┘
│
▼
┌─────────────────────────────┐ ┌─────────────────────────────┐
│ Self-Hosted API │ │ Edge Static Bundler │
│ FastAPI + PostgreSQL DB │ │ (service/edge_export.py) │
│ (service/ / Docker) │ └──────────────┬──────────────┘
└─────────────────────────────┘ │
▼
┌─────────────────────────────┐
│ Cloudflare Pages │
│ - dist/index.html │
│ - dist/data/feed.json │
│ - dist/data/descriptions/ │
└──────────────┬──────────────┘
│
▼
https://lilguy.win
- Client Feed (
dist/data/feed.json): Contains indexed metadata for all 8,300+ open opportunities, stripped of bulky HTML descriptions to keep network payload minimal. - On-Demand Description Shards (
dist/data/descriptions/<id>.json): Individual job descriptions are fetched on-demand when a user expands or inspects a posting. - Automated Discovery: Self-updating crawlers that verify and discover new ATS endpoints across companies.
To build the static distribution and deploy to Cloudflare Pages:
# 1. Export static edge bundle (generates dist/)
python3 -m service.edge_export
# 2. Deploy to Cloudflare Pages (requires wrangler)
./scripts/publish_edge.shTo preview locally:
cd dist && python3 -m http.server 8080
# Open http://localhost:8080# Launch database and background ingestion workers
docker compose up -d --build
# Open API docs at http://localhost:8000/docs
# Web interface available at http://localhost:8000# Install scraper dependencies
pip install -r scraper/requirements.txt
# Run full scrape against sources.yaml
python scraper/scrape.py
# Build a customized feed using a preset filter (e.g. Software Engineering)
python scraper/build_feed.py --filters presets/software-engineering.yaml --out SE_FEED.md| Directory / File | Description |
|---|---|
service/static/index.html |
Single-page client web app with responsive UI, search, and edge fallback |
service/edge_export.py |
Sharding and export pipeline for Cloudflare Pages edge distribution |
service/standardize.py |
Taxonomy normalization, location cleaning, and category classifiers |
scripts/publish_edge.sh |
Automated deployment script for lilguy.win via Wrangler |
data/all_postings.json |
Canonical raw dataset of all active internship opportunities |
sources.yaml |
Monitored company career sites and ATS endpoint configurations |
presets/ |
Ready-made filter definitions for major college majors and roles |
autofill/ |
Application autofill browser userscript and profile templates |
scraper/ |
Python scraping engine and vendor ATS connector drivers |
tests/ |
Pytest test suite covering normalization, categorization, and API routes |
Run the comprehensive test suite with pytest:
pip install -r requirements-dev.txt
pytest -v tests/Contributions are welcome! Please see CONTRIBUTING.md for guidelines on testing and verifying ATS connectors.
Distributed under the MIT License. See LICENSE for details.