01 · Roasts
Speed-Run Developer
Task-Manager: 4 commits in 5 minutes. File-Manager: created and done same day. Gold-Price-Prediction-Model: 2 commits in 85 seconds. You're not building software, you're doing time trials.
Secret Secret
MetroMart has 'metromart_secret_key' hardcoded directly in app.js. You built a supermarket management system with a loyalty points engine and 6 triggers — and then left the keys under the doormat.
README-Only Infrastructure
21 repos. 0 test suites. 0 CI pipelines. 0 licenses. The README flag is the only green light across your entire portfolio. You've mastered the art of documenting things that don't have tests.
Invisible to the Internet
0 followers, 5 total stars across 21 repos. Your portfolio site (Portfolio-PT) is built with GSAP animations and Nodemailer — but apparently nobody's received that email yet.
LSTM in 85 Seconds
Gold-Price-Prediction-Model: a full LSTM gold price predictor, committed start-to-finish in under two minutes. Either you're the fastest ML engineer alive or this is a copy-paste from a Kaggle notebook.
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% weight57D
- Depth15% weight50D
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
77 active days
Language distribution
- Jupyter Notebook34%
- JavaScript27%
- EJS13%
- Java9%
- HTML6%
- CSS5%
- Other6%
04 · Numbers
Owned repos
non-fork
16
Commits
last 12 months
202
Followers
0
Joined GitHub
Jul 2023
05 · Top repos
ParvTiwari /
Portfolio-PT
Personal portfolio built with Next.js 16, React 19, GSAP animations, and Tailwind CSS. Features contact form with email automation via Nodemailer and scroll-triggered animations. Lacks tests, CI, and license but demonstrates structured components and modern web practices.
ParvTiwari /
DocuSearch-AI
Personal RAG project built with Streamlit + Groq for document Q&A. Implements chunking and keyword-based retrieval without embeddings. Functional but lacks tests, CI, typing, and architectural documentation beyond README.
ParvTiwari /
MetroMart
Personal supermarket management system with Node/Express, Supabase, multi-table schema, but limited production signals—low stars (3), no tests, no CI, thin docs for non-DB code, hardcoded secrets.
ParvTiwari /
NewsMind-AI
Personal AI news platform with backend+frontend full stack built in ~1 hour; demonstrates typed JS, structured architecture, and working integration with NewsAPI+Groq LLM, but ultra-fresh (created 2026-03-28, 1 commit in 30s), no users or production signals.
ParvTiwari /
Shell
Educational Unix shell implementation in C with piping, redirection, and signal handling. Single-week effort (~4 commits in 6 minutes), 12 KB codebase, no tests/CI, demonstrable learning project with multiple OS concepts implemented.
ParvTiwari /
Task-Manager
One-off educational C project demonstrating basic Linux process management via system calls, minimal scope (5 KB, single file), created and completed within 5 minutes on 2026-04-19.
ParvTiwari /
File-Manager
Minimal educational C project demonstrating Unix file system calls. Single 500-line file with basic utilities (copy, merge, stat, cat, word count), no tests/CI, created and finished same day.
ParvTiwari /
Gold-Price-Prediction-Model
Single Jupyter notebook for gold price prediction using LSTM with minimal documentation, no tests, no CI, and only 2 commits created within 2 minutes. Essentially a one-off tutorial/exploration piece.
06 · Timeline
- Jul 10, 2023Joined GitHub
- Oct 9, 2025Created MetroMart
- Dec 30, 2025Created Portfolio-PT
- Feb 24, 2026Created Gold-Price-Prediction-Model
- Mar 5, 2026Created DocuSearch-AI
- Mar 28, 2026Created NewsMind-AI
- Apr 17, 2026Created Shell — This project is a lightweight implementation of a Unix-like command-line shell written in C. It supports core shell functionalities such as command execution, piping, input/output
- Apr 19, 2026Created File-Manager — A simple command-line File Manager built in C using low-level system calls. This project demonstrates core file handling operations in a Unix/Linux environment without relying on h
- Apr 19, 2026Created Task-Manager — A simple command-line based Task Manager implemented in C using Linux system calls. This project provides essential process management functionalities similar to basic system monit
- Apr 19, 2026Most recent push to Task-Manager
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.