LangGraph Agents — AI Job Research Assistant

A multi-agent pipeline built with LangGraph that researches a job title, extracts required skills, and produces a personalised gap analysis report against a candidate profile.

Built as a portfolio project demonstrating LangGraph state graph orchestration with Claude as the LLM backbone.

Docker Hub: hub.docker.com/r/omerfsen/langgraph-agents

Architecture

┌─────────────────────────────────────────────────────────────┐
│                    LangGraph StateGraph                     │
│                                                             │
│  ┌─────────────────┐     ┌──────────────────┐               │
│  │ fetch_job_reqs  │────▶│  extract_skills  │               │
│  └─────────────────┘     └──────────────────┘               │
│                                   │                         │
│                                   ▼                         │
│  ┌─────────────────┐     ┌──────────────────┐               │
│  │ generate_report │◀────│  gap_analysis    │               │
│  └─────────────────┘     └──────────────────┘               │
│           │                       ▲                         │
│           │               ┌───────────────┐                 │
│           │               │ load_profile  │                 │
│           │               └───────────────┘                 │
│           ▼                                                 │
│          END                                                │
└─────────────────────────────────────────────────────────────┘

Graph Nodes

Node	Responsibility
fetch_job_requirements	Uses Claude to generate realistic job requirements for the given title
extract_skills	Parses requirements into structured JSON (technical, soft, certs)
load_candidate_profile	Loads candidate profile from state (or uses default)
gap_analysis	Claude compares required skills vs profile → has/missing/partial
generate_report	Claude writes a markdown gap analysis report with recommendations

Shared State

class AgentState(TypedDict):
    job_title: str
    raw_job_data: Optional[str]
    required_skills: Optional[dict]
    candidate_profile: Optional[str]
    gap_analysis: Optional[dict]
    final_report: Optional[str]
    error: Optional[str]

Quick Start

git clone https://github.com/omerfsen/langgraph-agents
cd langgraph-agents

cp .env.example .env
# Add your ANTHROPIC_API_KEY to .env

pip install -r requirements.txt

python examples/run_example.py "AI Engineer"
python examples/run_example.py "Senior MLOps Engineer"

Docker

docker pull omerfsen/langgraph-agents:latest

docker run --rm \
  -e ANTHROPIC_API_KEY=your_key_here \
  omerfsen/langgraph-agents:latest

# Custom job title
docker run --rm \
  -e ANTHROPIC_API_KEY=your_key_here \
  omerfsen/langgraph-agents:latest \
  python examples/run_example.py "LLMOps Engineer"

Sample Output

==============================
  Job Research Agent: AI Engineer
==============================

[Step 1] Completed: fetch_job_requirements
[Step 2] Completed: extract_skills
[Step 3] Completed: load_candidate_profile
[Step 4] Completed: gap_analysis
[Step 5] Completed: generate_report

==============================
  GAP ANALYSIS REPORT
==============================

## Executive Summary
Strong infrastructure background with 85% skill match for AI Engineer roles...

## Strengths
- ✅ GPU infrastructure at scale (B200/H100 clusters)
- ✅ Kubernetes, Docker, Helm, Terraform
...

Tech Stack

• LangGraph — agent graph orchestration
• Anthropic Claude (claude-sonnet-4-6) — LLM for all reasoning nodes
• Python 3.14
• Docker — omerfsen/langgraph-agents

Related Articles

• MLOps vs LLMOps: Why Large Language Models Force Us to Rethink Production AI
• Build a RAG Pipeline Using Azure AI Search and Azure OpenAI with Terraform

Author

Omer Sen — Principal Consultant | DevOps · DevSecOps · MLOps · LLMOps