Simple LLM Agentic Job Application Assistant

The goal of the ChatBot is obtained the name, email, and skills from the user.

Demo of a Job Application Assistant using AI Agent:

High Level Explanation

The diagram of the states are shown below:

---
config:
  look: handDrawn
  theme: neutral
---
flowchart TB
    A["User Input"] --> B["LLM"]
    A -- \bye --> Exit[Exit]
    B --> C{"Need a tool?"}
    C -- No --> E["Generate Answer"]
    C -- Yes --> D["Request a tool"]
    E --> A
    D --> F["1. Extract Application Info
    2. Check Application Info"]
    F --> G["Use the selected tool and obtain information"]
    G --> B
    E --> DecisionExit{All info gathered?}
    DecisionExit -- yes ---> Exit
    DecisionExit -- no ---> A

Loading

The Infinite Loop will receive the user input and process it. If the user provided information in their prompt that is useful for the job application (name, email, skills) the tool extract application info will be called and save the values inside a dictionary. After that, the tool check application info is called and check if it is missing information and provide feedback to the user.

Note: The extract application info tool use Regex to extract the information. This can be improved using a more complex regex or using other LLM call.

Set-up

Python Libraries

• Only requires the litellm library for orchestration.

$ pip install litellm python-dotenv emojis

Google AI Studio

Note: You can use other model as OpenAI, Claude, ... You just need credits and an API Key. I'm using here Gemini because has a free tier.

1. First you need to generate a API KEY from Google AI Studio. Click on the button Get API Key.

Note: Don't share your API Key, keep it in secret.

Create .env file for your API Key

Create a file .env and paste your GEMINI_API_KEY generated before.

GEMINI_API_KEY=<your api key here>

Running

Be sure that your API KEY is working and run the main script

$ python app_job_application_assistant.py

---

✨ Crafted with passion by BrenoAV 🛠️