01 · Roasts
Sprint-and-Ghost Protocol
2,397 commits in a year across 83 repos, yet virtually every single one was created and 'completed' in under 24 hours. planetary-climate-dashboard, moisture-telemetry-mock, github-site-automator — born and abandoned on the same calendar day. You're not building software, you're generating GitHub activity.
Tests Are a Myth
Out of 12 repos scored, only ROIC-Intelligence-App and creditpulse-indian have actual test files. That's a 17% test coverage rate — on the repos, not even the code. retail-ledger-workstation has a CI pipeline that references pytest but no data to run it on. Aspirational CI is still just decoration.
The Streamlit Monoculture
Every single project is a Streamlit dashboard. Climate physics? Streamlit. IoT telemetry? Streamlit. AI deception research? Streamlit. Financial compliance? Also Streamlit. You've found your hammer and the entire universe is your nail.
0 Stars, 83 Repos
You've published 83 repositories and accumulated exactly 0 stars total. The math on that is brutal: infinite repos divided by zero community interest. Two followers, zero following, zero external PRs. GitHub is your personal journal and nobody has subscribed.
Hardcoded Credentials Speedrun
ai-portfolio-engine- ships with hardcoded API access blocks (literally 'srinivasta' gating in app.py), dummy random embeddings in sync.js, and malformed GitHub API URLs in github-site-automator. These aren't bugs — they're load-bearing placeholders being committed to public repos.
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% weight62C
- Consistency20% weight65C
- Quality20% weight38F
- Depth15% weight58D
- Breadth10% weight45D
- Community10% weight25F
03 · Stats
365-day commit heatmap
83 active days
Language distribution
- Python68%
- Jupyter Notebook25%
- HTML5%
- JavaScript2%
- PLpgSQL0%
- Dockerfile0%
04 · Numbers
Owned repos
non-fork
83
Commits
last 12 months
2,397
Followers
2
Joined GitHub
May 2024
05 · Top repos
SRINIVASTA /
creditpulse-indian
Early-stage Indian NBFC portfolio risk platform with RBI compliance focus. Typed Python, structured src/ layout, HAS_README + HAS_TESTS + HAS_CI, but experimental scope, minimal adoption (0 stars), and incomplete core application (app.py truncated mid-function).
SRINIVASTA /
ai-portfolio-engine-
Early-stage SaaS prototype converting GitHub profiles to AI portfolios with RAG chatbot. Styled Streamlit frontend, Node.js/Express backend skeleton, and Python ML pipeline. Typed language but lacks tests, CI, and has architectural gaps (dummy embeddings, hardcoded credentials, missing error handling).
SRINIVASTA /
my-grant-prototype
Personal prototype implementing a Streamlit dashboard with dual-layer AST scanning and XGBoost classification for software debt underwriting. Clean typed Python structure, but narrow scope, zero adoption signals, and thin test coverage limit impact.
SRINIVASTA /
ROIC-Intelligence-App
Early-stage financial analytics Streamlit app with Medallion Lakehouse architecture, PDF parsing, and Monte Carlo simulator. Typed language missing; has README, tests, CI, and documented structure but unproven adoption and modest codebase scope.
SRINIVASTA /
MATS_12_Financial_Deception_Probing_Gemma2
MATS application project studying deception detection in Gemma 2 via residual stream analysis. Typed-language lacking (Jupyter + Python), but structured with README, Streamlit dashboard (app.py), PyTorch notebook, and mathematical documentation. No tests/CI. Created and pushed same day (Aug 13, 2026).
SRINIVASTA /
gigo-xero
Personal project with working Streamlit + ML pipeline for financial SMS clustering, but thin type hints, no tests, suspicious licensing/tracking code, and lacks structured documentation beyond README.
SRINIVASTA /
planetary-climate-dashboard
New Streamlit+Scikit-Learn climate physics simulator ingesting NOAA telemetry with polynomial CO₂ forecasting and Random Forest classification, but created and completed in single day with zero adoption, no tests/CI, and untyped Python code.
SRINIVASTA /
moisture-telemetry-mock
Personal IoT telemetry dashboard prototype using Streamlit and Pandas. Real-time soil moisture simulation with dynamic thresholding, CSV logging, and UI controls. MIT licensed. No tests, CI, or type annotations.
SRINIVASTA /
retail-ledger-workstation
A one-week Python financial dashboard with CI but no README, untyped code, zero test coverage, and unverified business logic. Premature for production use.
SRINIVASTA /
github-site-automator
Streamlit app automating website generation via GitHub API with template system. No README, no tests, no CI, minimal documentation, and critical URL bugs (GitHub API endpoints malformed). Early-stage experiment with 28 of 30 recent commits in <24 hours.
SRINIVASTA /
creditpulse
One-shot experimental Streamlit app for RBI compliance checking on NBFC credit portfolios. Single entry point (app.py), minimal tests, no CI, no documentation beyond code comments. 165 KB codebase with 30 commits over 1 day suggests rapid prototype dump rather than sustained development.
SRINIVASTA /
shorts-trimmer-downloader
Single-day experimental Python project (71 KB) with no README, tests, CI, or documentation. Untyped code, no visible structure or examples, minimal public utility.
06 · Timeline
- May 27, 2024Joined GitHub
- Jun 26, 2026Created gigo-xero — An unsupervised automated bookkeeping data pipeline that cleans bank SMS text notifications, eliminates "Garbage In, Garbage Out" errors, clusters rows using K-Means, and streams a
- Jul 29, 2026Created my-grant-prototype — Automated codebase underwriting dashboard built with Streamlit and an embedded XGBoost classifier. Implements dual-layer static AST validation and classical ML inference to evaluat
- Jul 30, 2026Created ROIC-Intelligence-App — An AI financial simulator by Srinivasta tracking hyperscaler capital spend, multi-entity cost structures, and investment risk profiles using a cloud-native Medallion Lakehouse engi
- Aug 6, 2026Created creditpulse
- Aug 7, 2026Created creditpulse-indian — An automated portfolio risk optimization and card control engine for Indian NBFCs & Fintechs. Deploys an in-memory, zero-storage processing framework fully compliant with the RBI M
- Aug 8, 2026Created ai-portfolio-engine- — 🚀 Multi-tenant AI SaaS that automatically transforms public GitHub profiles into developer portfolios, featuring a secure, user-isolated RAG chatbot trained on repo documentation.
- Aug 10, 2026Created shorts-trimmer-downloader
- Aug 11, 2026Created github-site-automator
- Aug 13, 2026Created MATS_12_Financial_Deception_Probing_Gemma2 — An empirical study probing the residual stream vectors of google/gemma-2-2b-it during contrastive corporate finance compliance constraints. Features a native PyTorch layer-tracking
- Aug 14, 2026Created retail-ledger-workstation
- Aug 17, 2026Created moisture-telemetry-mock — Real-time IoT telemetry simulation dashboard built with Streamlit and Pandas featuring dynamic thresholding and automated control logic.
- Aug 19, 2026Created planetary-climate-dashboard — A live-streaming planetary climate physics simulation and machine learning dashboard driven by raw, real-time NOAA telemetry. Built with Python, Scikit-Learn, and Streamlit.
- Aug 19, 2026Most recent push to planetary-climate-dashboard
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.