01 · Roasts
Portfolio in overdrive
263 recent commit samples across 69 public repos: you ship like every idea has a deadline tomorrow.
Adoption loading
16 total stars and 1 fork across a portfolio this large means the engineering is outrunning the audience.
Guardrail collector
Finder and Vacant defend data floors, parser residue, PII, and snapshot collapse harder than many production teams defend production.
Hygiene roulette
relaybee brings strict TypeScript and adversarial CI, while FreeBee and Bus-Stop-Collisions forgot the tests-and-CI part of the checklist.
Built using
Zoral
Shadows one worker for a week, then takes over their job with zero extra setup. Behaves exactly like the original.
zoral.ai
02 · Category breakdown
- Impact25% weight68C
- Consistency20% weight80A
- Quality20% weight75B
- Depth15% weight58D
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
140 active days
Language distribution
- JavaScript43%
- TypeScript28%
- Python16%
- HTML8%
- CSS4%
- C++1%
04 · Numbers
Owned repos
non-fork
48
Commits
last 12 months
1,190
Followers
19
Joined GitHub
Oct 2022
05 · Top repos
EnesYilmazcode /
relaybee
Relaybee is a documented, typed TypeScript AI API proxy with encrypted provider-key blobs, OpenAI-compatible routing, opt-in supporter queues, security-focused tests, and deployed documentation at relaybee.vercel.app.
EnesYilmazcode /
Vacant
Vacant is a deployed, offline-capable Ohio State classroom-finder with substantial schedule harvesting, map/UI code, regression tests, and guarded CI despite zero stars and an untyped JavaScript implementation.
EnesYilmazcode /
Finder
Finder is a deployed Ohio State course-search tool with seat snapshots, instructor ratings, filters, calendar layout, deep links, and a carefully tested static-client architecture.
EnesYilmazcode /
MailSuiteFree
A focused Chrome MV3 extension with careful Mailsuite signature detection, beacon preservation, campaign-link safeguards, extensive jsdom coverage, and GitHub Actions testing; impact is currently limited by zero stars and no adoption evidence.
EnesYilmazcode /
Sparky
Sparky is a shipped browser-based 3D circuit designer with simulation, AI-assisted building, saving, and a live Render deployment; its low adoption is offset by substantial implementation scope and tests.
EnesYilmazcode /
Absent
A substantial, documented offline healthcare prototype combining FastAPI, FastSAM segmentation, local Gemma inference, geometry-based presence tracking, and a polished browser UI, but built as a one-day untested demo with known packaging and documentation gaps.
EnesYilmazcode /
Melody
Melody is a thoughtfully documented React/Vite offline music PWA with IndexedDB audio storage, playlists, lyrics, Media Session controls, and YouTube/a-Shell import flows, but has no visible adoption, tests, license, or typed code.
EnesYilmazcode /
ChessOverlay
A substantial local chess-screen recorder and Stockfish coaching suite with pixel recognition, overlays, PGN/JSON export, HoloChess, extensive headless checks, and CI, but no demonstrated adoption or typed implementation.
EnesYilmazcode /
nanoRAG
nanoRAG is a carefully engineered C++ exact-search foundation with aligned vector storage, deterministic top-k, dataset tooling, and focused regression tests, but it has no demonstrated adoption and the README is only a short project blurb.
EnesYilmazcode /
FreeBee
A polished, demo-oriented receipt-reward prototype with a phone UI, Flask relay, browser-use worker, survey clone, and Chrome extension, but currently showing no adoption and limited engineering validation.
EnesYilmazcode /
Bus-Stop-Collisions
Substantive NYC safety-analysis repo with documented BallTree/KNN workflows, multiple visualization scripts, and large source datasets, but no demonstrated adoption, tests, CI, typing, license, or reproducible engineering safeguards.
EnesYilmazcode /
ghostbuster
A focused, reproducible Python decoder with a FastAPI demo, synthetic encoder, README walkthrough, and pytest coverage, but currently a 0-star one-day project without CI, license, or typed code.
06 · Timeline
- Oct 4, 2022Joined GitHub
- Jan 19, 2025Created Bus-Stop-Collisions — Analyzing NYC bus stop crash risk and recommending safer relocations
- Feb 27, 2026Created Sparky — Agentic Circuit Designer
- May 27, 2026Created Melody — Offline Music Listening Service
- Jun 21, 2026Created nanoRAG — A from-scratch, SIMD-accelerated RAG retrieval engine in C++
- Jul 13, 2026Created ghostbuster — Decodes motion-masked fonts by tracking per-pixel velocity across frames.
- Jul 18, 2026Created FreeBee — Free Money
- Jul 29, 2026Created relaybee — Crowdsourced AI API
- Aug 1, 2026Created Absent — Nothing left behind
- Aug 8, 2026Created MailSuiteFree — Chrome extension that removes the Mailsuite/Mailtrack promotional signature from outgoing Gmail. Read tracking keeps working.
- Aug 19, 2026Created Finder — Find who teaches your classes before everyone else
- Aug 26, 2026Created Vacant — Find an empty classroom near you at Ohio State, free for as long as you need it. Knows which buildings are actually unlocked.
- Aug 31, 2026Created ChessOverlay — Optimal chess moves overlayed on any Chess platform
- Sep 6, 2026Most recent push to Finder
07 · Compare
08 · Rubric
How this score was produced
Overall = Σ (category × weight) + gentle top-end curve
Tier thresholds
▸ How the pipeline works
- 01Scrape.Pull every non-fork repo pushed in the last 90 days, plus your contribution calendar, followers, and language byte counts — straight from GitHub's REST & GraphQL APIs.
- 02Triage.A small model reads every repo's file tree + README and picks the 20 files per repo that actually reveal how you code.
- 03Grade each repo. All repos run in parallel through a fast scoring model that reads the picked files and rates each one independently on Impact, Quality, and Depth — with evidence citations.
- 04Aggregate. A larger reasoning model combines the per-repo scores with server-computed stats (heatmap, commit cadence, language entropy, follower count) to produce the 6-dimension profile score + roasts.
- 05Correct.Deterministic server-side checks enforce anchor-scale floors (e.g. a profile with 2,000+ public commits can't score 30 Consistency) and recompute the final verdict.
~90 seconds per profile, ~$0.25 in compute. Total of ~240 files read across your top-12 repos. One rating per GitHub account per day.
▸ Data sources & caveats
- Heatmap & commit totals: GitHub GraphQL
contributionsCollection— covers the last 365 days, includes private repos when the user has opted in (default). - Language %: byte totals across the top 30 owned non-fork repos.
- Curve: a small upward nudge centered on raw score ≈ 70, capping at 100. Prevents specialists from being unfairly penalised for narrow breadth.
- Anchor corrections: when server-measured signals (e.g. privateWorkLikely, multiRepoVolume, follower count) mandate a minimum category score, the aggregation step enforces it. These are signal-conditional, not identity-based floors.