Data Science student building AI agents, automation systems, and full-stack developer tools.
I like turning messy workflows into software — especially where LLMs, automation, data, and developer tooling intersect.
Portfolio · LinkedIn · X · Email · Dev.to · Substack
Terminal-native AI agent with multi-provider LLM failover.
Arka routes plain-English requests to local skills and tools while keeping the architecture provider-agnostic.
- 24 LLM providers
- 70+ built-in skills
- 75 MCP tools
- 320+ tests
- Local-first workflow
- Multi-provider failover
- CLI-native interface
- MCP support
- Automation-ready
pipx install "arka-agent[chat]"Links: GitHub · Documentation · PyPI
- LLM orchestration & provider failover
- RAG pipelines
- AI agents
- MCP integrations
- Computer vision
- OpenCV & MediaPipe
- Sentiment analysis
- Forecasting
- Whisper / TTS pipelines
- Browser automation
- Scheduled bots
- CLI tooling
- Desktop notifications
- Workflow automation
- Data extraction pipelines
- React
- Next.js
- TypeScript
- Tailwind CSS
- Framer Motion
- Vite
- REST APIs
- AI-powered web applications
- Docker
- Python packaging
- pipx
- PyPI
- CLI applications
- MCP servers/tools
- Bright Data
- Local-first developer workflows
Languages
Python TypeScript Rust JavaScript HTML/CSS
Frontend
React Next.js Tailwind CSS Framer Motion Vite
AI / ML
LLMs RAG MCP OpenCV MediaPipe Whisper TTS
Automation
Browser Automation Scheduled Bots CLI Tooling Desktop Automation
Infrastructure
Docker PyPI pipx Bright Data
AI Agents
↓
LLM Routing ─── MCP ─── Local Skills
↓
Automation ─── APIs ─── Data Pipelines
↓
Developer Tools ─── Full-Stack Apps
I'm particularly interested in systems that make AI useful, composable, local, and automatable rather than just another chatbot wrapper.
I write about things I'm building, experiments I'm running, and lessons from shipping software.
If you're working on AI agents, developer tools, automation, data products, or interesting open-source projects, I'm always interested in talking.
