01 · Roasts
Jupyter Notebook Hoarder
64% of your codebase is .ipynb files. Your GitHub profile is basically a collection of cells someone forgot to restart kernel on. Ship a package, not a notebook.
The 4-Hour Founder
Agentic-Web-Scrapper was conceived, born, and shipped in 4 hours across 7 commits on the same afternoon. At that velocity you could found and pivot a startup before lunch.
Zero Followers, 23 PRs
You opened 23 pull requests this year and somehow still have 0 followers. Either you're PRing your own repos in a loop or the GitHub algorithm has personally blacklisted you.
License? Never Heard of Her
FalconView, FalconMap, LoopIn, vaayu-ai — multiple repos with real engineering work and zero license files. Congrats, nobody can legally use your offline Dijkstra router.
Burst-and-Ghost Development Methodology
Area-Explorer: 5 commits, 1 day. FalconMap: 29 commits, 1 day. DataBros: 2 commits, 2 days. You have a very consistent pattern of very inconsistent follow-through.
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% weight65C
- Quality20% weight72B
- Depth15% weight58D
- Breadth10% weight72B
- Community10% weight28F
03 · Stats
365-day commit heatmap
136 active days
Language distribution
- Jupyter Notebook64%
- Dart8%
- Python8%
- JavaScript8%
- Java3%
- EJS2%
- Other7%
04 · Numbers
Owned repos
non-fork
25
Commits
last 12 months
466
Followers
0
Joined GitHub
Jul 2023
05 · Top repos
ParvTiwari /
vaayu-ai
Well-architected LangGraph multi-agent air-quality system with typed state, deterministic scoring agents, XGBoost forecasting, and rigorous separation of LLM (narration-only) from high-stakes decisions. README + tests + CI-free codebase shipped as a working ET AI Hackathon 2.0 submission.
ParvTiwari /
FloatNote
FloatNote is a real-time meeting intelligence desktop app combining live Whisper transcription, screen OCR, AI summaries via Groq, and a RAG chatbot—fully local-first with SQLite persistence. Well-architected Python backend (FastAPI, async, structured modules) paired with React 19 frontend and Electron wrapper, but lac
ParvTiwari /
FalconView
Tactical offline-first mapping app (Flutter, 106 MB) with cross-platform support, bundled MBTiles, offline routing, coordinate formats, and subscription tiers. Structured architecture, test coverage, and CI present; no license file.
ParvTiwari /
SHL
FastAPI agent for SHL assessment recommendation via Groq LLM, hybrid BM25+TF-IDF retrieval, and structured JSON responses. Personal take-home project, experimentally fresh, with schema validation and deterministic safeguards but minimal public adoption or production usage signals.
ParvTiwari /
MetroMart
Early-stage Express+EJS supermarket management system with robust PostgreSQL schema (schema.sql defines 13+ tables with triggers), functional CRUD routes, and Supabase integration. Minimal stars but shows non-trivial architectural work on a single named project.
ParvTiwari /
DineFlow
Personal restaurant management project with Express/Supabase backend, session-based auth, and cart/payment features. Typed language absent, tests missing, but documented README and structured multi-file layout meet quality floor. Single star and 1 fork indicate experimental/portfolio stage.
ParvTiwari /
DocuSearch-AI
Lightweight RAG document Q&A app using Streamlit + Groq API with keyword-based retrieval. Personal project with typed-adjacent Python, clear documentation, and working prototype; minimal adoption (1 star, no forks).
ParvTiwari /
DataBros
Brand-new Next.js 16 marketing site for Dataverse (custom database service) with TypeScript, Tailwind v4, form handling, SEO optimization. Personal project, ~2 days old, minimal adoption yet.
ParvTiwari /
Agentic-Web-Scrapper
Jupyter-notebook-based Streamlit web scraping RAG agent with LangGraph orchestration, claim verification, and confidence scoring. Young experimental project (4 hours old, 7 commits, 537 KB) using SentenceTransformers + FAISS + Groq LLM for cited research answers.
ParvTiwari /
FalconMap
Early-stage Android offline maps app with MapLibre, Nominatim integration, and Material 3 UI. Fresh project (created 2026-05-27) with clear README and typed Java foundation but no tests, CI, or license.
ParvTiwari /
LoopIn
Early-stage Flutter reminder app (3 recent commits) with core features (notifications, detection, DB) but incomplete documentation and minimal GitHub presence. Boilerplate README, no CI, no license, experimental connectors.
ParvTiwari /
Area-Explorer
Early-stage map/news exploration platform for Indian cities. Well-structured docs and roadmap, but only 5 commits in 1 day, untyped Python, no tests/CI, and minimal implementation beyond scaffolding.
06 · Timeline
- Jul 10, 2023Joined GitHub
- Jan 24, 2025Created DineFlow — A smart restaurant management system with real-time orders, admin dashboard, and scalable backend.
- Oct 9, 2025Created MetroMart — MetroMart is a complete solution for supermarket operations - from inventory management to fast billing and vendor procurement.
- Mar 5, 2026Created DocuSearch-AI — DocuSearch AI is an intelligent document Q&A system that allows users to upload PDFs, DOCX, or TXT files and receive instant AI-powered answers from their documents.
- Mar 28, 2026Created FloatNote — Real-time meeting intelligence — transcription, screen reading, AI summaries, and a chatbot that knows your meeting.
- May 17, 2026Created SHL
- May 27, 2026Created FalconView — Offline-first tactical mapping for Windows, macOS, Android, iOS & web — mark, track, measure and navigate, even off the grid. Built with Flutter & MapLibre.
- May 27, 2026Created FalconMap
- Jun 11, 2026Created Agentic-Web-Scrapper
- Jun 15, 2026Created Area-Explorer
- Jun 16, 2026Created LoopIn
- Jun 22, 2026Created DataBros
- Jul 2, 2026Created vaayu-ai — LangGraph multi-agent pipeline forecasting AQI 24–72h ahead (XGBoost), attributing pollution sources, and delivering guardrailed multilingual LLM advisories across 3 Indian cities.
- Aug 4, 2026Most recent push to DineFlow
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.