JobAssist

Your AI teammate for resume feedback, skill growth, and job discovery.

Comment

Inspiration

Despite the abundance of job boards like LinkedIn, Indeed, and SimplifyJobs, many job seekers—especially students and early-career professionals—still find themselves overwhelmed, overlooked, or underprepared. These platforms are reactive, not proactive: they bombard users with irrelevant roles, offer generic resume feedback, and rarely help candidates align their unique profiles with the right opportunities.

We saw a recurring pattern:

  1. Students applying blindly to hundreds of roles, unsure if their resumes are even being seen.

  2. Smart, skilled candidates losing out because they didn’t tailor their applications—or didn’t know how.

  3. Job boards offering quantity over quality, lacking meaningful insights or guidance.

  4. Ad-Boosted Noise on Existing Platforms

As job seekers ourselves, we knew there had to be a better way. We imagined an AI-powered career assistant—one that doesn't just list jobs but actively understands you, guides you, and works with you to land the right role. And that’s how JobAssist was born.

What it does

JobAssist uses the power of MongoDB's vector search and Google Gemini's AI capabilities to help users

  1. Upload their resume and receive AI-powered feedback

  2. Discover upskilling recommendations based on in-demand skills

  3. Get job role suggestions aligned with their experience and interests

  4. Leverage LLM embeddings to match resumes with job listings using vector search. More of sematic matching rather than just keyword matching.

  5. Save and track applied jobs within an intuitive UI.

  6. Generate cover letters and enhance resumes based on the job you are interested in.

How we built it

*For the demo site, we have scraped 400+ roles from job boards relevant to the field of data. (Roles like Data analyst, data scientist, data engineer, ML engineer, etc.) *

Frontend

  • Built with React + Vite,Tailwind CSS designed for responsiveness and clarity.

  • Deployed on Firebase Hosting

Backend

  • Data storage: MongoDB Atlas and GCS

  • FastAPI REST service containerized via Docker

  • Deployed to Google Cloud Run using CI/CD and secrets from Secret Manager

  • Startup handling via custom shell script with GCP credential bootstrapping

AI & Embeddings

  • Resume feedback and suggestions generated using Gemini 1.5 Pro

  • Embedding generation using Vertex AI Gemini Embedding model

  • Resume-to-job matching via MongoDB Atlas Vector Search

  • Job posting search using keywords—MongoDB Keyword Search.

Other Tools

  • Secrets: Google Secret Manager

  • IAM

Challenges we ran into:

  • LinkedIn actively blocks scraping, leading to missing information about the job postings.

  • Protobuf & gRPC version conflicts during Gemini integration

  • Google cloud service outage effecting deployments and APIs

  • Deciding on the embedding model and creating the needed token lengths.

  • Managing secure environment variables across services.

  • Designing an intuitive, non-overwhelming UI

Accomplishments that we're proud of

  • Built using RAG-style matching: resume & job embeddings + vector similarity search.

  • End-to-end deployment across GCP, Firebase, and MongoDB

  • Feedback system using generative AI

  • Secure, scalable, and modular architecture for production-readiness

  • Time & Team Coordination

What we learned

  • MongoDB's vector search (Embedding-based search) is a game-changer

  • How to orchestrate deployments on GCP.

  • Best practices in handling secrets and credentials at scale

  • Deepened understanding of embedding models and semantic search

  • Learned to balance UX, model latency, and cloud resource constraints

What's next for JobAssist?

  • Updated the database in real time with more information about the job roles.

  • Add Google OAuth for personalization

  • Resume version control and live editing with AI

  • LLM-powered career Q&A chatbot and interview prep materials.

  • Admin dashboard to track user insights and search behavior

  • Custom alerts and notifications

  • End-to-end customer personalization.

  • Build insights dashboard: application trends, job gaps, skill alerts.

Built With

Share this project:

Updates