yogeshhk/miningresume

Text Mining certain fields from a resume

Hi I'm Yogesh

I:

• 🌱 Create free educational content about Python, Machine/Deep Learning, etc. and open-source it under 'TeachingDataScience' repository here.
• 👯 Counsel for Mid-career transition to Data Science. My own transition is here and here is my TEDX talk on this topic. General Learning Path for such transition is here.
• 💬 Run office hours for any technical or career coaching queries, on Saturdays 2 to 5 pm (India time). Feel free to send email for appointment to my firstnamelastname at yahoo dot com . More details here and here
• 📫 Post on LinkedIn regarding AI-ML and allied topics. I would be very happy if you follow me there. 😊
• ⚡ Publish my blogs-stories on Medium as well. Feel free to follow me here.

Thank you 🙏

miningresume's Issues

Incorporate NER and Relationship Extraction

Study DBPedia neural extraction framework , past GSOC submissions and proposals, to devise good NER and Relationship Extraction

Other file formats

Add support to read pdfs and docxs

Extract keywords using Gen AI

Similar to regex used to mine resume, evaluate/implement the same by developing Gen AI

Extract keywords using Doc AI

Similar to regex used to mine resume, evaluate/implement the same by developing Doc AI resume processor

Parse resume programmatically using GPT or any other LLM models

Similar to regex used to mine resume, evaluate/implement the same by developing Prompts based resume parser

Custom NER

Build custom NER models using Spacy 3. Data from rijaraju repo is at data\rijaraju_repo_resume_ner_training_data.json

Refer

• https://www.youtube.com/watch?v=WpaioLNsoGI
• https://www.youtube.com/watch?v=YLQvVpCXpbU

Build KG

Based on the specifications, assume extractions have been done, then write code to build networkx graph. Visualize it.
Can it be visualized in Streamlit?

To Python 3.6+

Check if the script works in Python 3.6+ and if not make necessary changes

Building Web UI

Put flask UI on current python script
Add folder selection button, show results in tabular format