Technology Stack
Agentic AI Stack
Building autonomous agents, multi-agent systems, and agent frameworks. The frameworks and tools I use to build, experiment, and ship real solutions.
- BuildReal solutions
- ExperimentContinuously
- StayUp to date
- ShareWhat I learn
- Planning
- Tool Calling
- Memory
- Multi-Agent
- MCP
- Human-in-the-Loop
- Observability
Agentic AI
Building autonomous agents, multi-agent systems, and agent frameworks.
Agents that plan, use tools and act — built with guardrails so they're safe to run in production.
Overview
Agentic AI goes beyond chat: agents break goals into steps, call tools and APIs, keep memory and hand work to other agents. I build these systems with graph-based orchestration, standard tool protocols like MCP and clear human-in-the-loop checkpoints.
The goal is always reliability — agents that behave predictably, can be traced and evaluated, and fail safely.
Key Focus Areas
- Single and multi-agent architectures
- Tool calling and API integration
- Model Context Protocol (MCP) servers
- Agent memory and state management
- Human-in-the-loop workflows
- Tracing, evaluation and guardrails
Tools & Technologies
A selection of the tools and platforms I work with.
LangGraph
LangChain
CrewAI
- AutoGen
- Semantic Kernel
Pydantic AI
MCP
OpenAI
Claude
- LlamaIndex
n8n
FastAPI
Python
PostgreSQL
Redis
Docker
My Experience Level
Not just tools — real hands-on experience.
- LangGraph85%
- Multi-Agent Design80%
- Tool Calling & Function APIs90%
- MCP Servers80%
- CrewAI / AutoGen70%
- Agent Evaluation70%
Real-World Applications
How I use these technologies in practice.
- Research and analysis agents
- Customer-support agents with escalation
- DevOps and code-review agents
- Document-processing pipelines
- Agents connected to internal tools via MCP
- Workflow automation across SaaS apps
Featured Projects
- Research CrewA multi-agent system that conducts in-depth research on any topic using specialized agents (researcher, critic, summarizer).LangChainCrewAIOpenAI
- MCP IntegrationsExploring Model Context Protocol (MCP) with real-world tools like Google Drive, Notion, GitHub, and custom APIs.MCPPythonNotion API
- AegiSense AI AssistantAn AI-powered assistant for identity and security operations. Uses RAG, tool calling, and guardrails to answer security questions and automate workflows.LangGraphFastAPIPostgreSQL
What's Next?
Continuously learning and exploring.
- Long-running agents with durable state
- Better evaluation suites for agent behaviour
- More open-source MCP servers
- Agent-to-agent protocols
Learning Resources
Articles and tutorials I've written on these topics.
- TutorialBuild a secure RAG APIShip a RAG API where every answer respects who is asking with evals that catch regressions.
- Article · 15 min readAI Agent Architecture Patterns: Single Loop, Graph Workflow, Planner-Executor or Multi-Agent?Multi-agent designs inherit coordination, partial failure and unclear ownership from distributed systems, and they earn their place mainly when you need isolation. This guide compares four agent architectures on one realistic problem and ends with a framework for choosing.
- Article · 16 min readYour AI Agent Needs a Circuit Breaker: Bounding Loops, Tool Failures and Runaway CostWhen a tool starts failing, a naive agent keeps calling it until the run hits a limit or the bill arrives. A circuit breaker stops the loop, but only if it sits outside the model and trips on repeated calls, stalled progress and budget burn as well as on errors.
Explore topics:Agentic AI & MCP
Let's Build Something Amazing
Have a project in mind or want to collaborate? I'm always open to discussing new opportunities.
Let's Talk