01 · Roasts
Karpathy Cosplay
Two of your first repos are named 'makemore' and 'micrograd' — both empty or near-empty scaffolds cloned from Andrej Karpathy's work. At least one of them is 0 KB. The man built the originals in one sitting; you haven't written the first line.
README Millionaire, Code Pauper
ibcognito claims 10K+ MAU across 108 countries and 1M+ page visits — in a 3 KB README. The entire public repository is one markdown file. Either the most impressive README in history or the most optimistic one.
Sprint and Ghost
Out of 12 repos, at least 6 were created and last pushed within the same day or a 2-day window. makemore: 12 minutes. worldcup-markov-chains: literally 1 second between create and push. You start projects faster than most people name them.
Zero CI Across the Board
Not a single repo out of 15 has CI configured. Not hmm-world-cup with its test suite, not adaptive-fitness-intelligence with its 1.8 MB typed codebase, not ctc-mechanism targeting NeurIPS. GitHub Actions is free, Puranjay. It's right there.
The Heatmap Speaks
44 of 52 weeks show zero public commits. The burst in the last 5 weeks is real energy — but the 47 weeks of silence before it suggest your GitHub spent most of the year in cryosleep.
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% weight56D
- Consistency20% weight60C
- Quality20% weight57D
- Depth15% weight55D
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
21 active days
Language distribution
- TypeScript37%
- Python28%
- HTML20%
- CSS9%
- JavaScript5%
- R1%
04 · Numbers
Owned repos
non-fork
14
Commits
last 12 months
117
Followers
1
Joined GitHub
Apr 2024
05 · Top repos
puranjayh /
hmm-world-cup
Research-oriented Hidden Markov Model football predictor using Bayesian updating and Elo ratings. Typed Python codebase (50+ files implied by size), structured src layout with model/, scripts/, tests/, and evaluation pipeline. No CI/CD or license, but contains meaningful project documentation and working test suite.
puranjayh /
ctc-mechanism
Early-stage research project integrating scRNA-seq CTC datasets with R/Python scripts for transcriptomic analysis and immune evasion prediction. Project has meaningful scope but lacks tests, CI, type hints, and is currently incomplete.
puranjayh /
adaptive-fitness-intelligence
Typed React Native fitness tracking app with analytics. 1.8 MB codebase, 3 recent commits (2 days old), structured Expo project with multi-module architecture (types, helpers, analytics, API, screens, components). No tests, CI, or license; minimal README ("adaptive-fitness-intelligence" only).
puranjayh /
hypoxia-genomics-ai
ML pipeline for hypoxia gene classification with PCA, differential expression, and Random Forest (83.3% LOO-CV accuracy). Single-file analysis.py with comprehensive documentation but minimal commit depth (4 of 30 over 1 day), no tests/CI, untyped Python.
puranjayh /
black-scholes-pricer
Educational Black-Scholes options pricer implementing analytical BS formula, Greeks, and Monte Carlo from scratch without financial libraries. Well-documented with working math but minimal production scope and fresh creation.
puranjayh /
ibcognito
Student SaaS platform claiming 10K+ MAU across 108 countries with 1M+ page visits; WordPress-based with PHP backend. Public repo contains only README (3 KB); full source kept private. Claims content automation work but no code to evaluate.
puranjayh /
micrograd
A learning implementation of Andrej Karpathy's micrograd autodiff framework with basic Value, Neuron, Layer, and MLP classes. Minimal scope: 0 stars, 4 commits in 2 hours, no tests, no CI, no license, no gitignore. Raw educational code without quality infrastructure.
puranjayh /
puranjayh
Personal portfolio README listing educational background and research interests. 22KB repo with only metadata, no executable code, no tests, no CI, and no meaningful project artifacts.
puranjayh /
me
Personal HTML portfolio scaffold with zero stars, 0 forks, no tests, CI, or documentation; appears to be an initial commit-only project with minimal output.
puranjayh /
makemore
Empty scaffold with 0 stars, 0 files fetched, 1 of last 30 commits. No README, docs, tests, CI, license, or typed code. Created and last pushed 2026-05-28 (same day, ~12min apart).
puranjayh /
jarv
Empty scaffold with minimal README. Single commit on 2026-05-13, zero stars, no implementation files sampled, no tests or CI. Classic one-off dump.
puranjayh /
worldcup-markov-chains
Empty scaffold with no files, no documentation, and no meaningful commits. Repository created and immediately pushed with zero substance or implementation.
06 · Timeline
- Apr 17, 2024Joined GitHub
- Apr 18, 2026Created black-scholes-pricer — European options pricer built from scratch - analytical BS solution, all 5 Greeks, Monte Carlo via GBM. No financial libraries used.
- Apr 18, 2026Created hypoxia-genomics-ai — ML pipeline for classifying hypoxic vs normoxic cancer cell states from 60,000+ gene expression profiles - PCA, differential expression, volcano plots, Random Forest classifier. Bu
- May 1, 2026Created ctc-mechanism — R-script powered research on CTC evasion mechanism sc-RNA sq data
- May 10, 2026Created ibcognito — Global IB learning platform - 10K+ monthly users, 108 countries, 1M+ page visits.
- May 11, 2026Created puranjayh
- May 13, 2026Created jarv
- May 18, 2026Created worldcup-markov-chains — World cup predictions powered by Markov Chains
- May 23, 2026Created hmm-world-cup
- May 24, 2026Created me
- May 26, 2026Created adaptive-fitness-intelligence
- May 28, 2026Created micrograd — Andrej Karpathy's Micrograd
- May 28, 2026Created makemore — Andrej Karpathy's makemore
- May 28, 2026Most recent push to makemore
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.