01 · Roasts
Portfolio, meet verification
You have 231 public repos and 15+ named projects, yet minecraft, pg2hydra, and ai-interviewer still ship without CI or tests.
Commit volume is real
503 yearly commits and 117 cross-repo recent samples say builder; the uneven heatmap says your calendar occasionally takes unscheduled leave.
Agent factory mode
hyperagent has CI and serious tests; several neighboring AI-agent repos appear to have skipped that part of the assembly line.
Stars need a deployment pipeline
93 total stars across 231 public repos and only 4 forks: lots of shipping, limited evidence that strangers are carrying the boxes.
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% weight72B
- Depth15% weight55D
- Breadth10% weight65C
- Community10% weight55D
03 · Stats
365-day commit heatmap
200 active days
Language distribution
- TypeScript58%
- JavaScript23%
- Go13%
- CSS3%
- HTML2%
- Python1%
04 · Numbers
Owned repos
non-fork
43
Commits
last 12 months
503
Followers
39
Joined GitHub
Nov 2021
05 · Top repos
sundaram2021 /
hyperagent
A well-documented TypeScript/Python self-hosted AI agent platform with multi-provider streaming, MCP, memory, observability, encrypted secrets, sandboxing, and substantial automated coverage; currently a same-day, zero-star project.
sundaram2021 /
pg2hydra
A well-structured TypeScript PostgreSQL-to-HydraDB migration engine with introspection, relationship-aware rendering, batching, retries, uploads, and verification, but currently has no visible adoption, tests, CI, or license.
sundaram2021 /
inference-serving-system
A well-documented TypeScript inference-serving demo with dynamic batching, semantic caching, circuit breaking, observability, Docker Compose, and substantial tests, but it has 0 stars/forks and only a one-week history.
sundaram2021 /
ai-interviewer
A documented, typed Next.js voice-interview application integrating resume parsing, GitHub scraping, LLM/RAG interviewing, and Sarvam STT/TTS, with a structured multi-file implementation but no tests, CI, or license.
sundaram2021 /
minecraft
A substantial browser-based Three.js/React voxel sandbox with 23 WebMCP tools, procedural world generation, mobs, crafting, physics, and structure building, but currently has no visible adoption, tests, or CI.
sundaram2021 /
spaceinvader-vibe-jam
Starfall Armada is a documented, strictly typed Three.js/Vite arcade shooter with substantial gameplay scope, but it has no stars, tests, CI, license, or demonstrated external adoption.
sundaram2021 /
sturdy-funicular
A documented TypeScript LinkedIn profile API with structured normalization, session handling, caching, throttling, and a substantial Vitest suite, but no visible adoption, CI, or license and only a same-day commit history.
sundaram2021 /
sundaram2021
A maintained GitHub profile README documenting a broad portfolio of named AI and developer tools, with multiple live demos and domains, but no source files or repository engineering artifacts.
sundaram2021 /
sandbox-agent
A documented TypeScript sandbox-agent demo with a real HTTP/SSE agent loop, Anthropic tool orchestration, TensorLake isolation, and a polished browser UI, but no visible adoption, tests, CI, or license.
sundaram2021 /
observability-system
A documented, typed TypeScript/Express LLM observability demo with OpenTelemetry tracing and a browser dashboard, but only 1 star, no tests or CI, and a one-day two-commit history limit demonstrated adoption and depth.
sundaram2021 /
autonomous-browser-agent
A documented, typed TypeScript Playwright/Claude browser agent with modular observation, tool execution, sessions, overlays, and tabs, but it is a same-day zero-star project without tests, CI, or external adoption evidence.
sundaram2021 /
modiqo-hackathon
A newly created hackathon repository with only a minimal README and no source files, tests, CI, license, or meaningful implementation evidence.
06 · Timeline
- Nov 2, 2021Joined GitHub
- Feb 2, 2023Created sundaram2021
- May 1, 2026Created spaceinvader-vibe-jam
- Jun 9, 2026Created ai-interviewer
- Jun 12, 2026Created hyperagent — A Self Hosted HyperAgent Platform
- Jun 15, 2026Created observability-system
- Jun 16, 2026Created autonomous-browser-agent
- Jun 16, 2026Created sandbox-agent
- Jul 11, 2026Created inference-serving-system
- Aug 5, 2026Created pg2hydra
- Aug 31, 2026Created sturdy-funicular
- Sep 2, 2026Created minecraft
- Sep 4, 2026Created modiqo-hackathon
- Sep 4, 2026Most recent push to modiqo-hackathon
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.