Skip to content

Latest commit

 

History

History
210 lines (157 loc) · 8.58 KB

README.md

File metadata and controls

210 lines (157 loc) · 8.58 KB

Resumize (เรซูมั้ย?)

Overview

This application helps generate resume content based on your experience and the job description of your interest.

CleanShot 2567-04-28 at 01 25 21@2x

Article

Tech Stack

Features

  • Generate Resume Content: Input your dream job description and get a customized resume content. Please note that the result is not perfect.
  • Upload Experience: Upload your past experiences in JSON format to be used for generating the resume content.

Demo

Video demo

Demo - Resumize (Dioxus, Actix Web, Qdrant, LlamaEdge)

How it's designed

The project has been designed for few use-cases

  1. Upload my own experience (as JSON)
  2. Upload Job description from recruiter by copy from job market
  3. Create Resume content based on job description with suggestion

CleanShot 2567-04-28 at 01 25 38@2x

Upload my own experience (as JSON)

CleanShot 2567-04-28 at 01 26 00@2x

Description

The idea behind is I want to apply RAG on this project. So I think about the scenario that we can use to matching with resume content, Why don't we use some datasource or original resume? This decision as you already know, I was using dataset approach and it's easier for me to use JSON because I have it on my own portfolio site.

Here is current JSON format which acceptable by Backend

Reference: https://github.com/marttp/tpcoder-portfolio/blob/main/src/data/experiences.json

{
  "professional": [
    {
      "company": "company",
      "country": "country",
      "position": "position",
      "startDate": "startDate",
      "endDate": "endDate",
      "highlights": [
        "...list of achievement on this company"
      ]
    }
  ],
  "education": [
    {
      "school": "school",
      "degree": "degree",
      "fieldOfStudy": "fieldOfStudy",
      "startDate": "startDate",
      "endDate": "endDate",
      "courses": [
        "...list of string storing course"
      ]
    }
  ]
}

Enhancement

  • Consider to use PDF file should be general for many use-cases.
  • Could be implemented as asynchronous approach. User upload the file => Pass the file to workload which upsert document on vector database.
  • For each user upload, right now I read from constant value. However, I believe identify user identity on each upload should be much more usable on the future.

Upload job description and generate new resume content

CleanShot 2567-04-30 at 13 47 35@2x

Description

Find the role you interested, put job title and job description to get your suggestion and score of encouragement to apply your dream role depending on your past experience.

Enhancement

  • Currently, The result has shown in the group of text. Not the markdown. Meanwhile markdown result is responsed from LLM. I cannot do it on this project because currently, Dioxus don't have official Markdown support on desktop.
  • Because my Macbook Pro 2019 13 inches computation power could not process well with LLM project, I think it's better to designed as asynchronous approach as well as upload experience. For this scenarios, Send email about the result or send push notification and store it to inbox should be okay for the UX.
  • Better if the final result produce complete version of resume.

How to run locally

Prerequisite

Make it run locally

  • Clone the repository: git clone https://github.com/marttp/resumize

Run LlamaEdge for your local

Please follow this article about how to run WASM on your locally Getting Started with Llama-3-8B. Below are the steps required for the project only

  1. Install WasmEdge via the following command line.

    curl -sSf https://raw.githubusercontent.com/WasmEdge/WasmEdge/master/utils/install.sh | bash -s -- --plugin wasi_nn-ggml
  2. Download the Llama-3-8B model GGUF file.

    curl -LO https://huggingface.co/second-state/Llama-3-8B-Instruct-GGUF/resolve/main/Meta-Llama-3-8B-Instruct-Q5_K_M.gguf
  3. Download an API server app. It is also a cross-platform portable Wasm app that can run on many CPU and GPU devices.

    curl -LO https://github.com/LlamaEdge/LlamaEdge/releases/latest/download/llama-api-server.wasm
  4. Download the chatbot web UI to interact with the model with a chatbot UI.

    curl -LO https://github.com/LlamaEdge/chatbot-ui/releases/latest/download/chatbot-ui.tar.gz
    tar xzf chatbot-ui.tar.gz
    rm chatbot-ui.tar.gz
  5. Start an API server for the model

    wasmedge --dir .:. --nn-preload default:GGML:AUTO:Meta-Llama-3-8B-Instruct-Q5_K_M.gguf \
      llama-api-server.wasm \
      --prompt-template llama-3-chat \
      --ctx-size 4096 \
      --model-name Llama-3-8B

Feel free to take a look on the mentioned article if you couldn't understand them well. 🙇‍♂️

Run Qdrant on locally

Depending on your approach, I would suggest to using container technology to speed up on build your playground

Standalone container

  1. Open terminal and run below command if using docker - reference: https://qdrant.tech/documentation/quick-start/
    docker run -p 6333:6333 -p 6334:6334 \
    -v $(pwd)/qdrant_storage:/qdrant/storage:z \
    qdrant/qdrant

Compose file (I used this approach)

  1. Navigate to the project directory: cd actix-rag-qdrant
  2. Open terminal Run relate compose command (This example is using podman)
    podman compose up -d

Run Actix Web

  1. Navigate to the project directory: cd actix-rag-qdrant
  2. Run below command to start dev server
    cargo watch -x "run --bin server"

Run Dioxus

Locally

  1. Navigate to the project directory: cd dx-app
  2. Install npm: https://docs.npmjs.com/downloading-and-installing-node-js-and-npm
  3. Install the tailwind css cli: https://tailwindcss.com/docs/installation
  4. Run the following command in the root of the project to start the tailwind CSS compiler:
    npx tailwindcss -i ./input.css -o ./assets/tailwind.css --watch
  5. Run the following command in the root of the project to start the Dioxus dev server:
    dx serve --hot-reload --platform desktop

Bundle version (Only for macOS)

  1. Follow all step from Locally section

  2. Stop development mode and run below command instead

    dx bundle --release
  3. If you success, You will have this dmg file.

    CleanShot 2567-04-30 at 21 03 36@2x

Apologize for haven't tried on Windows or Linux, I have only 1 laptop. 🙇‍♂️

License

This project is licensed under the MIT License.

Contact

If you have any questions, feel free to reach out to me at below contact