Motivation

Applying for jobs requires tailoring your resume and cover letter to each role — a repetitive process that's easy to skip, even though it meaningfully improves your odds. The goal of this project is to make tailoring effortless: the app knows your full experience history and the job you're targeting, and an LLM does the drafting.

The project has two technical motivations beyond the use case. First, I wanted to run a generative AI pipeline locally using Ollama — combining clever information retrieval with prompt engineering rather than relying on a cloud API. Second, I wanted to demonstrate using AI coding tools (Claude Code) to deliver a working product quickly in areas I'm less familiar with, like frontend development.

Current Progress

The frontend and backend are being developed in parallel, guided by up-front design specs for the database schema and API contract. The current build supports:

  • Experience tracking — store roles, responsibilities, projects, and accomplishments in a structured database
  • Application tracking — log job postings with company, role, and status
  • LLM integration — a local model (via Ollama) drafts resume and cover letter content given the job description and retrieved experience

The LLM integration is currently the weakest link: running a small local model produces output that needs more editing than I'd like. This is expected — the architecture intentionally uses an OpenAI-compatible API (Ollama's endpoint), which means swapping to a cloud model is a one-line config change when the time comes. The same applies to the database layer, which was designed with migration in mind.

What's Next

Development is phased so each stage is independently testable before the next begins:

  1. Phase 1 — Database review: Audit the schema and operations for correctness and edge cases
  2. Phase 2 — Information retrieval: Add vector search over experience entries; LLM-assisted company research to give the model better context on each employer
  3. Phase 3 — Resume generation: Iterate on prompts for resume and cover letter drafting; evaluate output quality systematically
  4. Phase 4 — Frontend: Polish the UI after the backend is fully validated — cleaner experience tracking forms, better application dashboard
← All Projects