I'm Aditya Pathak (also found here), a Graduate Trainee Engineer/Gen AI Consultant at Cubastion Consulting.
I joined the Vatika Business Park office in Gurgaon on January 2nd, 2025.
S. No. | Project Name | Project Description |
---|---|---|
1 | Company Database APIs | This project uses Python (Flask-SQLAlchemy), PostgreSQL (PgAdmin 4) and Postman (for API Testing) to enable CRUD functionality and more for a database 'Company', with many tables and constraints in those tables. |
2 | faceMatching | This project, developed with VS Code, Jupyter Notebook and Google Colab, uses Python (Flask, Pytorch, face_recognition, and more) and Postman (for API Testing) to develop two implementations of recognizing human faces, particularly those present in the LFW dataset and Indian Actors Dataset (both available on Kaggle). |
3 | Milvus Querying | This project uses Python, Hugging Face (sentence-transformers), Milvus + Docker (container running Vector DB) to create a vector database, populate it with details of many people (names, ages, salaries, addresses and their introductions) and enable searching and querying on the database contents using Cosine-Similarity distances on IVF Flat index. |
4 | Cubastion HR Chatbot | Presenting, Cubastion's HR chatbot - it can answer queries based on all the latest HR documents published by Cubastion's HR team. This conveniently saves time, allowing a Cubastion employee to resolve their query without having to comb through the actual documents. Developed with Python, sentence-transformers, Pinecone, llama3.2, and Streamlit. |
- Joined the GEN AI - FUSO team
- Co-developed the TIC Chatbot for Mitsubishi FUSO
- Performed comprehensive testing of the application by querying and comparing expected responses
S. No. | Project Name | Project Description |
---|---|---|
1 | Match Resumes with Job Description | This project aims to make the process of matching resumes with a particular job description much faster. Simply enter the required job-description and all the resumes that need to be filtered and run the script to find the top scorer as well as the 'n' best matching resumes! Built using Python, Hugging Face and Scikit-Learn. |
2 | AI Agent 01 | This is my first AI Agent. It can chat, perform basic calculations, reverse strings, set timers, and summarise text. Currently a work in progress. |
3 | iLikePDF | Open source tools that give great results, like the well-made PDF handling service, iLovePDF. There are - pipelines for converting directories of PDFs to DOCX and PPTX files. - APIs to convert single PDFs to DOCX and PPTX files effortlessly. Currently a work in progress. |
Mayank Sharma and I began R&D work on a 'Smart HR Chatbot' system, that uses our own, custom-built version of an Agentic AI system to perform many repetitive, mundane and time-consuming tasks, simply with a conversational interface which does the heavy lifting in the background. This application is currently a work in progress.
At this moment, not many details can be shared about this project, but here are a few functionalities that have been implemented as of 30th May, 2025:
- Answering queries regarding Cubastion Consulting's HR Policies (by Retrieval Augmented Generation using HR Policy documents available on xNet).
- Referring someone in the user's network for a job. If required, the agent can also show a list of currently open positions in the company.
- Suggest holidays based on the user's leave requests and company-recognized holidays, and submit leave applications on their behalf.
- (As a manager) Get feedbacks for the user's team members, compare their performances, and even submit new feedbacks for them.
- Display the status/balances of the user's benefits, such as leave balance and holiday calendar.
- Or simply tell the user a quote to brighten up their day!
- Python and it's libraries, such as
Langchain
,Streamlit
, and many more - AzureOpenAI -
gpt-4o
for the LLM - Pinecone vector database
- MySQL databases
- Linux VM for hosting