# Vishesh Baghel - Portfolio

> Software engineer specializing in AI agents, vector systems, and full-stack development. Building production-ready integrations with TypeScript, React, Next.js, and modern AI tooling. Open source contributor to Mastra.ai and other frameworks.

I work with TypeScript, React, Next.js, Node.js, Spring Boot, PostgreSQL, and vector databases. I contribute to open source frameworks and take on selective freelance work—shipping MVPs, hardening infrastructure, and integrating AI into real products.

## About

- [About Me](https://visheshbaghel.com): Overview of skills, experience, and current work
- [Client Work](https://visheshbaghel.com/work): Past projects and collaborations
- [Open Source Contributions](https://github.com/vishesh-baghel): Active contributions to Mastra.ai and other frameworks
- [Contact](mailto:visheshbaghel99@gmail.com): Get in touch for consulting or collaboration

## Experiments & Technical Writing

### Optimizations
- [LLM Router - Intelligent Model Selection](https://visheshbaghel.com/experiments/llm-router-cost-optimization): Standalone LLM router utility with ML classification, semantic caching, streaming, and multi-provider support. Tested with 200 training examples and 100 test queries.

## MCP Server

- [Experiments MCP Server](https://visheshbaghel.com/mcp): Access production-ready code patterns directly in your IDE
- [NPM Package](https://www.npmjs.com/package/vishesh-experiments): Install via npx for Cursor, Windsurf, or Claude

## Technical Stack

- **Languages**: TypeScript, JavaScript, Java
- **Frontend**: React, Next.js, HTML/CSS, TailwindCSS
- **Backend**: Node.js, Express.js, Spring Boot, Microservices
- **Databases**: PostgreSQL, MySQL, MongoDB, Redis, Vector DBs (Milvus, LanceDB, ConvexDB)
- **Cloud & DevOps**: AWS (Lambda, SQS, S3, RDS, Cognito), Docker, Git
- **AI Tooling**: Mastra, Vector databases, AI agent development, LLM pipelines
- **Testing**: Jest, Vitest, JUnit5, Mockito, TDD/BDD

## Projects

- [Kestra Agent](https://github.com/vishesh-baghel/kestra-agent): AI-powered workflow builder converting natural language to Kestra YAML
- [Glidee Bot](https://github.com/vishesh-baghel/glide): AI code review assistant with 85% risk prediction accuracy
- [Docs Chatbot](https://github.com/vishesh-baghel/docs-chatbot): Deploy chatbots that understand your documentation
- [Kestra MCP Server](https://github.com/vishesh-baghel/kestra-mcp-doc-server): MCP documentation server utilities

## Optional

- [LinkedIn](https://www.linkedin.com/in/vishesh-baghel/): Professional profile and connections
- [Twitter/X](https://x.com/VisheshBaghell): Updates and technical insights
- [Calendar](https://cal.com/vishesh-baghel/15min): Book a 15-minute consultation