01 · Roasts
Notebook monoculture
92% of the code mix is Jupyter Notebook: strong lab energy, but the production surface area is still mostly cells.
Zero-user launch
Three named projects, 0 total stars, 0 forks, and 1 follower: the portfolio shipped, but nobody has boarded yet.
Automation allergy
Sanyam-Ag, NNTutorial, and CalHelpr all report no CI and no automated tests. The bugs are currently enjoying sovereign immunity.
Sprint, not saga
CalHelpr has 10+ modules, but its sampled 14 commits landed across a 5-day age window—promising build, unproven maintenance arc.
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% weight23F
- Consistency20% weight35F
- Quality20% weight39F
- Depth15% weight35F
- Breadth10% weight55D
- Community10% weight25F
03 · Stats
365-day commit heatmap
41 active days
Language distribution
- Jupyter Notebook92%
- Python6%
- C++1%
- JavaScript0%
- HTML0%
- CSS0%
- Other1%
04 · Numbers
Owned repos
non-fork
7
Commits
last 12 months
83
Followers
1
Joined GitHub
Jul 2025
05 · Top repos
Sanyam-Ag /
CalHelpr-CallForHelp
A substantial local-first Python benefits navigator with document parsing, deterministic FPL calculations, program matching, Streamlit workflow, deadline tracking, and follow-up chat, but no tests, CI, license, or demonstrated adoption.
Sanyam-Ag /
NNTutorial
An educational ML-from-scratch repository spanning micrograd, classical regression/classification, and character-level makemore notebooks, with meaningful implementation depth but no demonstrated adoption, tests, CI, or packaging.
Sanyam-Ag /
Sanyam-Ag
A polished GitHub profile README for an ML/AI identity, but with no demonstrated software artifacts, adoption, tests, CI, or typed implementation in the supplied repository evidence.
06 · Timeline
- Jul 10, 2025Joined GitHub
- Aug 16, 2025Created Sanyam-Ag — Hi, thanks for visiting, contact me if you're interested in collaborating with me :)
- May 13, 2026Created NNTutorial — Tweaking Around NN to get some inspiration
- Jun 16, 2026Created CalHelpr-CallForHelp — Governmental assistance program matcher that builds your profile using your docs and chat in natural language using local SLM and then matches to the programs with an application a
- Sep 23, 2026Most recent push to Sanyam-Ag
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.