How devanand17tech-ai/aijobsearch works
This is a Streamlit AI job-search agent. The user submits a PDF and search preferences; the agent extracts a candidate profile, plans queries, searches live Apify listings or demo data, normalizes and deduplicates results, matches and ranks jobs, then presents recommendations and apply links. The graph follows documented workflow and visible imports/calls. Some agent, UI, and LLM implementation details are unsampled, so those connections rely on README documentation or are omitted where evidence is insufficient.
Main entry points: tools.py Agent Tools · agent.py Search Agent · client.py LLM Client · app.py Streamlit Dashboard
Components 18
User Interface
- Streamlit Dashboard
app.py
Agent Orchestration
- Search Agent
agent/agent.py - Execution State
agent/state.py - Agent Tools
agent/tools.py - Preference Memory
memory/memory.py
Job Discovery
- Job Search Manager
jobs/job_search.py - Apify Job Scraper
jobs/apify_client.py - Job Normalizer
jobs/normalizer.py
Profile and Matching
- Resume Parser
resume/parser.py - LLM Client
llm/client.py - Profile Prompt
prompts/resume_prompt.py - Search Prompt
prompts/search_prompt.py - Match Prompt
prompts/matching_prompt.py - Match Evaluator
matching/matcher.py - Result Ranker
matching/ranker.py
Other components
- Job Seeker
- Live Job Listings
- Job Recommendations
Connections 22
- Job Seeker → Streamlit Dashboard: uploads and configures
README.md - Streamlit Dashboard → Search Agent: runs search
app.py - Streamlit Dashboard → Preference Memory: loads preferences
app.py - Search Agent → Agent Tools: uses tools
agent/agent.py - Search Agent → Execution State: records progress
agent/agent.py - Search Agent → Preference Memory: saves preferences
agent/agent.py - Agent Tools → Resume Parser: extracts text
agent/tools.py - Agent Tools → LLM Client: profiles and plans
agent/tools.py - LLM Client → Profile Prompt: formats profile prompt
llm/client.py - LLM Client → Search Prompt: formats search prompt
llm/client.py - LLM Client → Match Prompt: formats match prompt
llm/client.py - Agent Tools → Job Search Manager: searches jobs
agent/tools.py - Job Search Manager → Apify Job Scraper: queries scraper
jobs/job_search.py - Apify Job Scraper → Live Job Listings: fetches listings
jobs/apify_client.py - Job Search Manager → Job Normalizer: normalizes and deduplicates
jobs/job_search.py - Job Search Manager → Agent Tools: returns listings
agent/tools.py - Agent Tools → Match Evaluator: evaluates fit
agent/tools.py - Match Evaluator → LLM Client: requests rationale
matching/matcher.py - Agent Tools → Result Ranker: ranks and filters
agent/tools.py - Search Agent → Job Recommendations: produces results
README.md - Streamlit Dashboard → Job Recommendations: presents job cards
app.py - Job Recommendations → Job Seeker: offers apply links
README.md