LEPA Support Agent
A LangGraph support workflow for LEPA: typed state, conditional routing, tools, memory, and a multi-agent split behind FastAPI.
- Python
- LangGraph
- FastAPI
- SQLite
- OpenRouter
AI Learning Tracker
A public repo of small AI engineering projects — terminal agents, MCP, LangGraph, and async document workflows — built while learning in public.
Overview
AI Learning Tracker is my public learning and project repo. Personal study logs stay private; the code is meant to be cloned. Phase 1 is TAM, DevAssist, and an agent harness. Later projects add a LangGraph LEPA support agent, a Temporal document pipeline, and an AI job application assistant with human review.
Problem
I wanted to move into AI/LLM engineering by shipping small, complete projects instead of notebook demos — and to keep the code public.
What I built
Each project teaches one idea: tool calling, MCP, a thin harness, then a LangGraph support agent, durable document orchestration, and a job-application workflow that stops for a human. The repo is the index; articles on Medium walk through what I learned.
My contribution
- Shipped TAM, DevAssist, the agent harness, LEPA Support Agent, a document pipeline, and a job assistant as sequential public projects.
- Documented each project in public write-ups.
Technologies
- Python
- TypeScript
- OpenRouter
- MCP
- AI SDK
More ai / llm
AI Document Processing Pipeline
An async document pipeline: upload or webhook, Temporal extract/analyze/summarize, then Trigger.dev email — without blocking the request.
- Temporal
- Trigger.dev
- TypeScript
- Webhooks
- LLM
AI Job Application Assistant
Parse a job and a CV, explain the match, draft materials, then stop for a human. The AI never applies.
- Python
- LangGraph
- FastAPI
- Human-in-the-loop
- Evaluation
Related writing
Agentic AI Is a Workflow, Not a Smarter Prompt
From reflection and tool use to evals, planning, and multi-agent systems — the concepts that changed how I think about building AI applications.
I Built an AI Job Application Assistant
Parse a job and a CV, explain the match, draft materials, then stop for a human. The AI never applies.
I Built an AI Document Processing Pipeline
How I combined webhooks, Temporal workflows, AI summarization, async notifications, and email into one small system.
What I Learned About Async AI Workflows, Webhooks, and Orchestration
Jobs, events, durable workflows, and background tasks — the concepts I needed before building an AI document pipeline.
Putting It All Together: The LEPA Support Agent
What Project 2 became — end to end — and what I would change next.
Keeping a Human in the Loop
What HITL actually means — and why asking “which LEPA area?” is not the same thing.
Breaking LEPA Support Agent Into Multiple Agents
Intake, knowledge, and support — why split the work, and how LangGraph keeps the specialists in order.
Giving LEPA Support Agent Memory
Checkpoints, thread IDs, and how a LangGraph support agent remembers a conversation across turns.
Giving LEPA Support Agent Tools
Why agents need tools, how search_knowledge works over Markdown docs, and how the knowledge node uses them.
Building LEPA Support Agent with LangGraph
Project setup, state, nodes, edges, and how a support request actually moves from START to END.
What I Learned Before Building My First LangGraph Agent
How I finally made sense of orchestration, state, memory, and multi-agent workflows — and why I decided to build a LEPA support agent next.
I Built a Local AI Terminal Agent With Tool Calling (And You Can Too)
How I built TAM — a CLI assistant that lists files, searches code, runs shell commands, and talks to free models on OpenRouter.