01 · Roasts
Speed-run Developer
12 repos and barely a single one has more than one day of commits. git_reverse (2026-04-02 to 2026-04-03), Hackathon_CBC (2.5-hour burst), AutoEDA (same-day dump) — you're not building projects, you're speedrunning repo creation.
Test? Never Heard of Her
0 out of 12 repos have tests. Not one. HAS_TESTS=no across the entire portfolio. You've written LLM judge pipelines, EDA agents, and health AI systems, but apparently unit tests are someone else's problem.
README Lottery
4 of your repos have no README at all — including Jarvis, Mood_Detection, and Benchmark_SLM. Mood_Detection is literally an empty folder. At least Startup_Eval has a README; it's just the default Vite template one.
2 Followers, 6 LLM Providers
git_reverse supports Gemini, OpenAI, Anthropic, Grok, OpenRouter, AND Hugging Face — yet has 1 fork and 2 followers. You've integrated more AI providers than you have GitHub followers.
IIT Delhi Builder™
Bio says 'Builder, Entrepreneur, Innovator' — and you do have 12 repos in 6 months, which is genuinely prolific. But with 51 public commits, 0 PRs, 0 issues, and 94% solo work, the building is happening very quietly.
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% weight56D
- Consistency20% weight55D
- Quality20% weight57D
- Depth15% weight35F
- Breadth10% weight65C
- Community10% weight25F
03 · Stats
365-day commit heatmap
23 active days
Language distribution
- Python52%
- JavaScript20%
- TypeScript14%
- CSS8%
- HTML6%
- Java0%
04 · Numbers
Owned repos
non-fork
15
Commits
last 12 months
51
Followers
2
Joined GitHub
Oct 2024
05 · Top repos
bruce249 /
Hackathon_CBC
VitalWeave: a Streamlit health data AI assistant integrating Luna Ring wearables, mood logs, and food diaries via Claude API for multi-signal pattern detection and personalized health insights. Personal project with ~193KB codebase, 10 recent commits, no tests/CI.
bruce249 /
AutoEDA
Modular multi-agent system for automated EDA with ingestion, analysis, modeling, and chat agents. Typed Python + structured documentation + FastAPI backend, but no tests/CI, zero stars, created 3 days ago with 4 recent commits.
bruce249 /
Small_LM_Benchmark
Personal evaluation benchmark project with structured async pipeline, typed Python, comprehensive README, and multi-agent architecture. Early-stage burst of 9 commits across 4 days with no tests or CI.
bruce249 /
gpt-advanced
Feature-rich React+Vite AI chat application with multi-provider support (OpenAI, Gemini, Ollama, etc.), voice I/O, document chat, and Mini GPT explanations. Early-stage personal project with solid architectural structure and typed libraries, but minimal stars/adoption and sparse commit history.
bruce249 /
git_reverse
Chrome Extension using React 19 + TS to analyze GitHub repos with 6 LLM providers (Gemini, OpenAI, Anthropic, Grok, OpenRouter, Hugging Face) and local model support. Modular llm-service.ts with provider adapters, github.ts file tree filtering, storage.ts for secure API key management. Typed, documented, structured cod
bruce249 /
TeamBrain
Personal Next.js project with AI-powered team knowledge management, featuring TypeScript, Prisma DB, NextAuth, and OpenRouter integration. 4 commits over 4 minutes indicates fresh, experimental codebase with minimal real-world adoption.
bruce249 /
Model_Council
Early-stage debate framework orchestrating three LLMs with judge scoring and verifier pipeline (math, citations, code, logic). 3 stars, no license, single day of commits, minimal architectural foundation.
bruce249 /
bruce249.github.io
Personal portfolio site built with Next.js + TypeScript, featuring animated landing page and command palette. Minimal substance: 72 KB, created/pushed same day, placeholder pages, no tests or docs.
bruce249 /
Benchmark_SLM
Fresh dump of an LLM benchmarking API backend—single-day burst with 0 stars, no README/tests/CI. Unpolished foundation with typed FastAPI + SQLAlchemy but no production readiness.
bruce249 /
Jarvis
One-shot initial dump of a PyQt6 AI desktop assistant scaffold. Zero stars, created today, 1 commit in 75 seconds. No README, tests, CI, license, or gitignore. Python code is untyped, unstructured, and untested.
bruce249 /
Startup_Eval
Fresh Vite+React boilerplate scaffold with zero commits beyond creation. Untyped, untested, no CI, generic README—a one-off empty starter template.
bruce249 /
Mood_Detection
Empty scaffold with no files, no README, no documentation, and zero commits since creation. No discernible project substance.
06 · Timeline
- Oct 27, 2024Joined GitHub
- Feb 11, 2026Created gpt-advanced
- Mar 4, 2026Created Small_LM_Benchmark
- Mar 8, 2026Created AutoEDA
- Mar 13, 2026Created Jarvis
- Mar 18, 2026Created Mood_Detection
- Mar 24, 2026Created Startup_Eval
- Mar 27, 2026Created Benchmark_SLM
- Mar 28, 2026Created TeamBrain
- Apr 2, 2026Created git_reverse
- Apr 8, 2026Created bruce249.github.io
- Apr 12, 2026Created Hackathon_CBC
- Apr 19, 2026Created Model_Council
- Apr 19, 2026Most recent push to Model_Council
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.