01 · Roasts
98% Jupyter, 0% Production
Your language breakdown is 98% Jupyter Notebook. You're not building software — you're building slide decks that occasionally run cells. CodeConclave is the only repo that saw actual JavaScript.
AADHAAR: 160 Files, 0 Working Endpoints
The AADHAAR-PULSE-SYSTEM has 160+ Python files, an IMPLEMENTATION_PLAN.md, and a Streamlit app that is literally half a function. Planning-to-shipping ratio: ∞.
quant-learning: 2 Commits, 3 Minutes, Big Dreams
Your HFT learning repo has 2 commits timestamped 23:25 and 23:28 on the same night. That's not learning — that's a 3-minute existential crisis with a .md extension.
CI? Never Heard of Her
Across 7 repos scored, HAS_CI=yes appears exactly 0 times. GitHub Actions has been sitting there the whole time, free of charge, judging you silently.
CodeConclave Carrying the Whole Roster
12 of your 16 total stars, the only tests, the only license, the only ARCHITECTURE.md — CodeConclave is doing Atlas-level heavy lifting while your other 32 repos take the day off.
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% weight30F
- Community10% weight40D
03 · Stats
365-day commit heatmap
114 active days
Language distribution
- Jupyter Notebook98%
- Python1%
- JavaScript0%
- Java0%
- HTML0%
- Shell0%
- Other1%
04 · Numbers
Owned repos
non-fork
21
Commits
last 12 months
107
Followers
15
Joined GitHub
May 2021
05 · Top repos
lukiod /
CodeConclave
A React-based code editor with Monaco integration, terminal (xterm), and Jupyter notebook support. Typed with styled-components, documented, and structured; ships with HAS_README, HAS_TESTS, HAS_LICENSE. ~2.2 MB codebase with multi-layer architecture (auth, editor, file system, code execution). 30 of last 30 commits ov
lukiod /
classification-pipeline
Image classification pipeline with Hydra config, TensorFlow/Keras CNN, class weighting, and W&B logging. Typed Python, documented structure, but minimal adoption (0 stars), no tests, no CI, experimental scope.
lukiod /
CoderLion
GitHub PR reviewer agent with FastAPI backend and Next.js frontend using Gemini API; typed Python with multi-agent architecture but lacks tests, CI, and polished documentation.
lukiod /
AADHAAR-PULSE-SYSTEM---UIDAI-HACK
Jupyter-based UIDAI hackathon project analyzing Aadhaar enrollment patterns. Sized ~80MB with structured src/ modules (anomaly detection, forecasting, preprocessing), but lacks tests, CI, license, and executable delivery. Quality hindered by incomplete implementations, no working example notebooks, and Jupyter-only for
lukiod /
ecgnet
Educational ECG classification project built in Jupyter notebooks with deep learning and ML baselines. No tests, CI, license, or type checking. Primarily a coursework artifact with promising results (97.94% CNN accuracy) but minimal structural integrity and production readiness.
lukiod /
lukiod
Personal GitHub profile README (59 KB) with no source code, tests, or CI. Contains only social badges and GitHub stats widgets. No meaningful project implementation or documentation.
lukiod /
quant-learning
Personal learning journal/checklist for HFT education created Feb 9, 2026 with only 2 commits in ~3 minutes. Minimal structured content—mostly unchecked checkbox lists and incomplete notes. No code, no tests, no CI, no license.
06 · Timeline
- May 28, 2021Joined GitHub
- May 25, 2024Created lukiod
- Feb 28, 2025Created CodeConclave — A powerful, AI-enhanced code editor that supports multiple programming languages with real-time syntax highlighting, intelligent autocompletion, and seamless debugging. Designed fo
- Jun 24, 2025Created ecgnet — This project focuses on ECG beat classification to distinguish between normal and abnormal heartbeats using deep learning techniques. It involves preprocessing ECG signals, extract
- Sep 15, 2025Created CoderLion — a code reviewer agent which replaces coderrabbit with a gemini api
- Jan 9, 2026Created AADHAAR-PULSE-SYSTEM---UIDAI-HACK
- Feb 5, 2026Created classification-pipeline — This is a default classification pipeline which can be used for testing stuff by adding the data changing the config changing the model
- Feb 9, 2026Created quant-learning — my weekly learning of quant and notes of it
- Apr 26, 2026Most recent push to classification-pipeline
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.