01 · Roasts
Portfolio, not audience
Five named projects are shipping, but 5 total stars says the internet has barely found the showroom.
CI is a rare Pokémon
arcer has CI; the other four scored repos leave verification to optimism and local machines.
Tests by technicality
bpho owns the authoritative test flag, while epe-video-editor's 13 fixture shapes still do not translate into a recognized test setup.
Deepest rabbit hole
smc-sorter lasted roughly eleven months and cleared 30 commits; the rest of the portfolio is still auditioning for long-term maintenance.
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% weight48D
- Consistency20% weight55D
- Quality20% weight52D
- Depth15% weight50D
- Breadth10% weight55D
- Community10% weight40D
03 · Stats
365-day commit heatmap
128 active days
Language distribution
- Python71%
- HTML28%
- JavaScript1%
- TeX1%
- CSS0%
- Shell0%
04 · Numbers
Owned repos
non-fork
19
Commits
last 12 months
145
Followers
13
Joined GitHub
Aug 2025
05 · Top repos
kavan-vyas /
natwest-fiducia
A substantial local FastAPI/Ollama credit-risk prototype with deterministic scoring, SQLite audit storage, reports, CSV export, and a polished HTML interface, but currently has no visible adoption, tests, CI, or license.
kavan-vyas /
epe-video-editor
A polished, documented local lesson-video assembler with a substantial ffmpeg pipeline and browser UI, but only 1 star and no demonstrated external adoption, CI, license, or typed implementation.
kavan-vyas /
smc-sorter
A documented but lightly adopted utility that pairs question/answer GIFs and generates OneNote-friendly PDFs; main.py has useful structure and type annotations, but requirements, README, and implementation diverge and there are no tests or CI.
kavan-vyas /
arcer
Documented arcer scaffold for a macOS arm64 Chromium patch workflow, with disciplined shell tooling and CI, but no patches or implemented browser features yet and only 1 star.
kavan-vyas /
bpho
A zero-star, one-shot computational-challenge repository combining exploratory Python scripts with MATLAB nuclear-physics calculations, but lacking documentation, CI, and a coherent tested package structure.
06 · Timeline
- Aug 24, 2025Joined GitHub
- Sep 6, 2025Created smc-sorter — sorts smc questions
- Nov 9, 2025Created epe-video-editor — Video editor created for Eleven Plus Exams
- May 15, 2026Created arcer
- Jul 6, 2026Created natwest-fiducia — This is the Fiducia Credit score calculator app
- Jul 26, 2026Created bpho — My main attempt 2026 at bpho computational challenege
- Aug 19, 2026Most recent push to smc-sorter
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.