01 · Roasts
Ghost Committer
totalCommitsYear = 0 on the public heatmap — 52 weeks of solid white. Either you're doing all your work in private repos or your keyboard takes more breaks than you do.
Burst-and-Ghost Engineer
radio_temp: 28 commits in 3 days. Recall: 24 commits in 3 days. AllStateJarvis: 5 commits in 2 days then abandoned. Your entire GitHub is a series of hypomanic weekends followed by radio silence.
80% Notebooks, 0% Tests
Jupyter Notebook is 80% of your codebase and you have exactly one repo with any tests at all. Those aren't codebases — they're lab notebooks you forgot to submit.
0 Followers, 0 Following, 0 PRs
You've been on GitHub since June 2022, built 29 repos, and have literally zero followers, zero following, and zero external PRs. You're developing in a hermetically sealed vacuum.
README Halfway House
OralScan's README ends mid-sentence. AllStateJarvis has no README at all. Recall has ARCHITECTURE.md but no README. You write documentation the way people write grocery lists — halfway through and then you remember you forgot milk.
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% weight60C
- Quality20% weight72B
- Depth15% weight65C
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
0 active days
Language distribution
- Jupyter Notebook80%
- Dart11%
- Python3%
- C++2%
- TypeScript2%
- C1%
- Other1%
04 · Numbers
Owned repos
non-fork
29
Commits
last 12 months
0
Followers
0
Joined GitHub
Jun 2022
05 · Top repos
VeeraSaiJoshik /
radio_temp
TypeScript + Python radiology copilot with Electron UI, Gemini Live integration, and Django/FastAPI backend. Typed, tested, CI-free, shipped product with structured architecture spanning 116KB and 28 recent commits.
VeeraSaiJoshik /
Halo
Early-stage Flutter desktop trading app with embedded WebView, market data polling, and multi-tab architecture. Untyped Dart with structured codebase, meaningful documentation, and 30+ commits in ~3 weeks but minimal production adoption.
VeeraSaiJoshik /
Recall
Early-stage competitive beat-recreation game with TypeScript Next.js frontend, Web Audio sequencer, and Firebase backend. Active development (24/30 commits in past month) with structured codebase but no tests, CI, or production polish. Personal experimental project.
VeeraSaiJoshik /
OralScan
Early-stage oral cancer detection system combining multi-modal ML pipeline (U-Net segmentation, InceptionV3 classification) with cross-platform Flutter mobile app and FastAPI backend. Demonstrates technical scope but lacks production infrastructure, tests, and CI/CD.
VeeraSaiJoshik /
LLM-Token-Compression-Middleware
Personal experiment in LLM token optimization using NLP and embeddings. Has README + typed Python, modular structure with 6 optimizer strategies (TOON, compression, cache, whitespace, semantic cache, deduplication), but incomplete implementation (hanging CLI code, unfinished async functions), no tests/CI, unpolished.
VeeraSaiJoshik /
AllStateJarvis
Early-stage competitive programming template library with Textual TUI; project incomplete with no README, tests, CI, or license. Repository shows experimental structure and abandon.
VeeraSaiJoshik /
Group_Chat_Jeapordy
Early-stage experimental transformer-based Jeopardy game generator from group chat messages. Python project with minimal documentation, no tests/CI, incomplete implementation (stub returns []), but demonstrates coherent NLP pipeline design across multiple modules.
06 · Timeline
- Jun 9, 2022Joined GitHub
- Mar 20, 2025Created OralScan
- Mar 9, 2026Created Group_Chat_Jeapordy — A transformer powered platform to automatically create a Jeapordy set based on the Group Chat's wildest and most unhinged messages!
- Mar 11, 2026Created LLM-Token-Compression-Middleware — Novel caveman adaptation that aims to reduce LLM token bloat while maintaining accuracy using classical NLP methods ensuring free and fast compression.
- Mar 12, 2026Created radio_temp
- Mar 30, 2026Created AllStateJarvis
- Apr 3, 2026Created Recall
- Apr 6, 2026Created Halo — The browser for day traders
- Apr 28, 2026Most recent push to Halo
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.