Find the edge
Look for concrete issues: deprecated APIs, encoding bugs, async loop mistakes, schema mismatches, and missing coverage.
Hello, I'm
I build practical automation, AI tools, and developer workflows, backed by 195 PRs opened, 59 merged PRs, and contributions to high-impact repos totaling 1138k+ stars.
Self-taught developer from the Philippines. I got into programming because I was tired of doing repetitive things by hand — so I automated them. That became building AI systems, contributing to open-source projects with tens of thousands of stars, and running multi-agent LLM pipelines on local hardware.
I have opened 195 PRs and contributed across 30 repositories, with 59 merged PRs in production codebases. The work spans bug fixes, tests, docs, refactors, and developer tooling across projects like pydantic-ai, chroma, flowsint, langchain, llama_index, mem0, and agno.
I use AI-assisted workflows to build and ship faster — all projects are designed, tested, and maintained by me.
def get_developer():
return {
"name": "Rolly Calma",
"stack": ["Python", "Flask", "Docker"],
"prs_opened": 195,
"merged_prs": 59,
"gpu": "local LLMs ✓",
"status": "open_to_work",
}
I borrowed this from good design tools: show the work while it is happening. My best PRs are small enough to review quickly, but complete enough to include the bug, the fix, and the verification trail.
Look for concrete issues: deprecated APIs, encoding bugs, async loop mistakes, schema mismatches, and missing coverage.
Change the smallest surface that fixes the behavior, matching the repo style instead of inventing a new pattern.
Run focused checks, compile touched modules, and explain why broader tests were or were not needed.
Thank maintainers, handle bot process requests, and follow up calmly when CI or review finds something real.
Scripts, pipelines, and bots that run unattended — from file processing workflows to multi-step API integrations.
Multi-agent systems, local LLM setups, and practical AI tools using Ollama and Claude API — production-ready, not just demos.
Flask backends, REST APIs, and real-time streaming apps with SSE — from working prototype to deployed and running.
195 PRs opened across 30 contributed repos, including 59 merged fixes, tests, docs, and refactors. Small, practical changes that make maintained projects easier to use.
Hire me for done-for-you automation — or inspect the tools and sample outputs before you decide.
Payment↔invoice matching, AR aging, drafted collection emails, and duplicate-bill detection — from a CSV export. For small accounting firms & SMBs. Free proof run.
View AR/AP service →Find & fix the SEO, mobile, accessibility, and broken-link issues costing you traffic and trust. Prioritized report, 48-hour delivery. Free mini-audit.
View Webloom service →Paste any URL and get an instant A–F website report — SEO, meta tags, social-share preview, and accessibility checks. Plus a real AR/AP sample report and a live Mini-RAG demo with retrieval + evals + guardrails.
Run the live auditor → AR/AP report Try Mini-RAGFree host — the first load may take ~30s to wake.
Short proof pages for the deployed demos and local AI tools reviewers should inspect first.
A deployed URL analyzer that turns SEO, accessibility, and page-health checks into a client-ready launch report.
Read case study →A compact retrieval demo built around visible source grounding, guardrails, and lightweight evaluation checks.
Read case study →A Flask multi-agent workspace with meeting presets, saved memory, streaming, and transcript export.
Read case study →15 zero-dependency Python modules, each drop-in ready. Covers thread-safe rate limiting, retry with full-jitter backoff, structured logging, async helpers, and HMAC signing — tested in workflows handling hundreds of API calls per session. Zero external dependencies, pure stdlib only.
A public shelf of small runnable tools under GhravenLabs: Data Cleaner, WPM Trainer, Website QA Auditor, Mini-RAG, and Steam Free Games Tracker. Each repo includes a screenshot, case study, smoke-check workflow, and clear README so reviewers can inspect proof quickly.
A polished Flask multi-agent workspace where local Ollama agents and optional cloud advisors join by @mention. Recent product upgrades added Code Review, Product Debate, Research, and Planning templates, searchable saved memory, Markdown transcript export, customizable agent personas, free-talk SSE streaming, and a portable Windows .exe release. Built as a practical meeting tool, not just a demo chat app.
A compact retrieval-augmented generation demo with document loading, semantic search, answer grounding, and lightweight evaluation checks. Built to show the practical pieces employers and clients ask about: chunking, retrieval quality, guardrails, and clear failure notes instead of a black-box chatbot.
Two battle-tested utilities pulled from production code. A thread-safe token-bucket RateLimiter supporting configurable call budgets per second, plus a @retry decorator with full-jitter exponential backoff — handles hundreds of API requests per session reliably. Zero dependencies, drop-in for any project.
I'm open to junior Python roles, AI automation work, open-source collaboration, and practical tools that need someone persistent enough to ship.