ResuMate: AI Career Assistant
LLM-powered career assistant for job preparation
ResuMate - AI Powered Career Assistant
Project Overview
ResuMate is an LLM-powered application designed to help users prepare for job interviews and optimize their career documents. It is especially useful for students about to graduate, recent graduates seeking jobs or internships, and working professionals looking to enhance their portfolios. Powered by the Gemma 2 9B model from GroqCloud, ResuMate offers a seamless experience by enabling all features with just a resume upload.
Key Features
- Chat with your resume
- Generate a good looking portfolio
- Craft a personalised resume
- Craft a job-personalized cover letter
- Search for jobs
- Interview Simulator
These features help users create tailored resumes and cover letters for specific jobs, generate clean and professional portfolio websites (publishable via Vercel or Netlify), and search for job opportunities. The interview simulator is especially beneficial for students to improve their skills and confidence for job interviews.
Technology Stack
Learnings & Takeaways
- Developed primarily using GitHub Copilot Agent Mode, leveraging expertise in Python and Streamlit
- Solves real-world problems and encourages brainstorming for new features
- Improved understanding of Python-based frontend tools like Streamlit
- Motivated to build more AI-powered projects
Live Demo
Dual Character Fusion: Creating a Merged AI Portrait
AI image generation using ComfyUI and LoRA models
Dual Character Fusion: Merged AI Portrait
Project Overview
This project explores image diffusion using ComfyUI, FLUX Dev, FLUX Dev LoRA, and AI-Toolkit. At its core, it involves generating a single image by merging the visual characteristics of two individuals—myself and my mother—through individually trained LoRA models.
Workflow Overview
- LoRA Training: I trained two LoRA models—one on my own images and one on my mother's—using a PyTorch instance in JarvisLabs. The trained models were then hosted on Hugging Face for seamless access.
- Base Image Selection & Workflow Setup: I selected a mother-son photo as the base reference and constructed a Depth ControlNet workflow. This was combined with a FLUX-style generation path, initially loaded with my LoRA.
-
Segmentation & Fusion: Using
Florence2RunandSam2Segmentation, I isolated one subject from the base image. The masked section was then routed through a second branch where my mother's LoRA was applied. Finally, I used theDetailernode to refine the fused image, blending both styles into a realistic and cohesive portrait.
Learnings & Takeaways
This project deepened my understanding of:
- The inner workings of image generation and transformation using AI
- Training LoRA models on custom datasets
- Deploying cloud GPUs for high-performance model training and experimentation
Ultimately, I created personalized, high-quality AI-generated images that beautifully merge identities—making it not just a technical experiment but a meaningful creative endeavor.
ASEB Seating Arrangement
Automated exam seating arrangement system for universities
ASEB Seating Arrangement
Project Overview
A specialized seating arrangement generator for Amrita Vishwa Vidyapeetham, Bengaluru Campus. This Python-based automation tool streamlines the exam seating arrangement process by generating comprehensive PDFs for classroom layouts, attendance sheets, and student guides. It also sends automated email notifications to students about their exam locations.
The system efficiently manages seating arrangements for all B-Tech courses, including 💻 Computer Science (CSE), 🤖 Artificial Intelligence (AIE) , 📊 Artificial Intelligence and Data Science (AID) , 📡 Electronics & Communication (ECE) , ⚡ Electrical & Electronics (EEE) , ⚡💻 Electronics & Computer (EAC) , ⚡📡 Electrical & Computer (ELC) , 🛠️ Mechanical (MEE) , and 🤖 Robotics & Artificial Intelligence (RAE) Engineering programs.
Learnings & Takeaways
- Developed with GitHub Copilot assistance, enhancing both Python coding and prompting skills
- Gained valuable experience working with various LLMs including Claude Sonnet, ChatGPT, and Gemini
- Automated the entire exam seating arrangement process, saving significant administrative time
- Implemented automated email notifications for better student communication
- Created comprehensive PDF generation system for classroom management