01 · Roasts
Shipping, not yet spreading
You have 12 named projects and 867 yearly commits, but the portfolio has only 10 stars total. The code is traveling farther than the audience.
CI knows the good stuff
queuemaxxing has tests, race checks, vet, and CI; several otherwise serious projects still stop at “works on my machine.”
Horizontal boss fight
271 multi-repo recent commits says you build broadly. Now pick Bartie or Queuemaxxing and make one repo impossible to ignore.
Demo allergic reaction
me_me_me has WebGL polish but its Vercel homepage returned DEPLOYMENT_NOT_FOUND; image-classification-transfer-learning serves hard-coded probabilities.
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% weight79B
- Depth15% weight58D
- Breadth10% weight80A
- Community10% weight55D
03 · Stats
365-day commit heatmap
146 active days
Language distribution
- Python45%
- TypeScript29%
- Jupyter Notebook8%
- JavaScript7%
- Go3%
- Rust3%
- Other5%
04 · Numbers
Owned repos
non-fork
40
Commits
last 12 months
867
Followers
21
Joined GitHub
Feb 2022
05 · Top repos
BABTUNA /
Bartie
A documented, typed Go CDC engine with PostgreSQL WAL capture, Kafka buffering, snapshot backfill, idempotent MERGE writes, crash tests, and focused unit coverage; adoption is not yet evidenced by its 0 stars and 0 forks.
BABTUNA /
queuemaxxing
A thoughtfully engineered Go durable HTTP queue with WAL recovery, priority/FIFO/LIFO/delay semantics, strong concurrency and persistence tests, and CI; it is an early single-week project with no demonstrated adoption.
BABTUNA /
carbonpilot
CarbonPilot is a documented FastAPI routing prototype with CI, pytest coverage, modular routing/cache/batching components, and a dashboard-oriented architecture, but it has no visible adoption, license, or typed implementation.
BABTUNA /
me_me_me
A polished, typed Next.js 15 personal portfolio/blog with MDX content, accessibility touches, and substantial WebGL presentation, but no tests, CI, license, or demonstrated external adoption.
BABTUNA /
quietroute
QuietRoute is a documented FastAPI/Next.js accessibility-aware routing prototype with PostGIS schema, multi-objective routing sketches, CI, and focused tests, but it has no visible adoption and several implementations remain minimal.
BABTUNA /
image-classification-transfer-learning
Documented PyTorch transfer-learning study with CNN/ResNet18 models, imbalance handling, Grad-CAM utilities, tests, and a browser demo, but no demonstrated adoption, CI, license, or real checkpoint-backed demo.
BABTUNA /
code-four-video-search
A substantial, documented body-camera video search system with a multi-stage retrieval and VLM verification pipeline, meaningful unit tests, and strong engineering design, but no demonstrated adoption and no CI or typed implementation.
BABTUNA /
multidomain-sentiment-transformers
Documented educational sentiment-analysis study with modular classical and transformer workflows, reproducibility notes, paper artifacts, and focused tests, but no demonstrated adoption, CI, license, or retained experimental outputs.
BABTUNA /
ergo-takehome
A substantial, documented take-home delivering a four-channel synthetic deal-graph pipeline, evidence-backed detectors, cached LLM extraction, and a React/Vite replayable UI, but with no adoption signals, CI, license, or typed Python.
BABTUNA /
BABTUNA
A personal GitHub profile repository centered on README presentation and automated 3D contribution graphics, with little substantive product or library code.
BABTUNA /
system-design-interview-prep
A focused system-design interview study guide with a ranked 16-question curriculum and an unusually detailed Bitly design, but no implementation, tests, CI, or adoption signals.
BABTUNA /
turbo-remote-cache-enabled-repro
A narrowly scoped Turborepo configuration reproduction documenting an apparent environment-variable wiring bug, with a minimal pnpm workspace and no implementation, tests, or CI.
06 · Timeline
- Feb 7, 2022Joined GitHub
- Apr 6, 2024Created multidomain-sentiment-transformers — Final Project for COP 4530 : Data Structures
- Apr 28, 2024Created image-classification-transfer-learning
- May 2, 2024Created quietroute
- Aug 11, 2024Created BABTUNA
- May 5, 2025Created carbonpilot
- May 22, 2026Created me_me_me — all about meeeeeeeeeeeeeeee
- Aug 21, 2026Created turbo-remote-cache-enabled-repro — Minimal repro: TURBO_REMOTE_CACHE_ENABLED is never wired into turbo config
- Aug 22, 2026Created queuemaxxing — A durable HTTP queue supporting FIFO/LIFO, priority, and delayed delivery
- Aug 22, 2026Created system-design-interview-prep — Structured system design interview study plan and detailed question breakdowns
- Aug 28, 2026Created code-four-video-search
- Sep 6, 2026Created ergo-takehome
- Sep 9, 2026Created Bartie — A small, correct change data capture engine: Postgres WAL to a destination database with backfill, idempotent merges, and crash recovery. Built as a study of streaming replication.
- Sep 17, 2026Most recent push to BABTUNA
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.