Skip to content

User Profile & Target Criteria

Jump to bottom
Chazona Baum edited this page Jun 24, 2026 · 2 revisions

Relevant source files

The Profile subsystem is the "North Star" of the Lodestar application. It encapsulates the user's professional identity, career history, and specific job-seeking preferences. This data is stored entirely within the profile/ directory of the Obsidian vault as Markdown files and is consumed by the pipeline to perform high-recall filtering and high-precision fit scoring.

Overview

The profile system serves two primary functions:

  1. Identity & Context: Providing a rich dataset of accomplishments, experience, and community involvement to the LLM for qualitative alignment analysis.
  2. Targeting & Scoring: Defining the "ideal" role through quantitative criteria (salary floors, levels, locations) and weights that drive the automated scoring engine.

System Architecture

The following diagram illustrates how profile data flows from the vault into the core logic of the application.

Profile Data Flow

flowchart LR
    subgraph Consumers
        Filter["pipeline/filter.rs"]
        Scoring["fit.rs (Scoring Engine)"]
        LLM["prompts.rs (Alignment)"]
    end
    subgraph subGraph1 ["Rust Entities (src-tauri/src/profile.rs)"]
        TargetCriteria["struct TargetCriteria"]
        Accomplishment["struct Accomplishment"]
        Experience["struct Experience"]
        Community["struct Community"]
    end
    subgraph subGraph0 ["Vault (profile/)"]
        TC["target_criteria.md"]
        ACC["accomplishments/*.md"]
        EXP["experience/*.md"]
        POS["positioning.md"]
        COMM["community/*.md"]
    end
    TC --> TargetCriteria
    ACC --> Accomplishment
    EXP --> Experience
    POS --> LLM
    COMM --> Community
    TargetCriteria --> Filter
    TargetCriteria --> Scoring
    Accomplishment --> LLM
    Experience --> LLM
Loading

Sources: src-tauri/src/profile.rs#1-172src-tauri/src/pipeline/filter.rs#1-43

Profile Data: Accomplishments, Experience & Community

The user's professional background is distributed across several sub-directories. The system uses a standard pattern of reading all non-template Markdown files in a directory, splitting the YAML frontmatter from the body, and stripping wikilinks from tags.

  • Accomplishments: Short notes in profile/accomplishments/ that map specific achievements to competency slugs.
  • Experience: Chronological work history in profile/experience/. The system calculates total_months_experience to drive seniority scoring.
  • Community: Volunteer and community roles in profile/community/ used to provide a holistic view of the candidate.
  • Positioning: A narrative "elevator pitch" in profile/positioning.md.

For details on parsing and the experience envelope calculation, see Profile Data: Accomplishments, Experience & Community.

Sources: src-tauri/src/profile.rs#125-161src-tauri/src/community.rs#1-49

Target Criteria & Fit Weights

The target_criteria.md file defines the boundaries of the job search. It is parsed into the TargetCriteria struct src-tauri/src/profile.rs#37-56

Field Purpose
match_titles A list of keywords used by the prefilter to identify relevant jobs during scraping src-tauri/src/pipeline/filter.rs#19-23
comp_floor The absolute minimum salary required; used as a "dealbreaker" in scoring.
fit_weights A set of five integers summing to 100 that prioritize different dimensions of a role src-tauri/src/profile.rs#15-22

The FitWeights struct allows users to tune the engine. For example, a user focused on salary over tech stack would increase the comp weight and decrease the skills weight.

For details on targeting fields and weight distribution, see Target Criteria & Fit Weights.

Sources: src-tauri/src/profile.rs#15-56src-tauri/src/pipeline/filter.rs#31-43

Fit Scoring Engine

The Fit Scoring Engine is the primary consumer of TargetCriteria. It transforms a raw Job entity into a FitBreakdown by comparing the job's attributes against the user's profile.

Scoring Logic Bridge

flowchart LR
    Final["Weighted Total (0-100)"]
    subgraph subGraph1 ["Scoring Dimensions"]
        S["Seniority Score"]
        SK["Skills Score"]
        C["Comp Score"]
        A["Arrangement Score"]
        D["Domain Score"]
    end
    subgraph subGraph0 ["Code Entities"]
        TC["TargetCriteria (profile.rs)"]
        J["Job (job.rs)"]
        FE["score_fit (fit.rs)"]
    end
    TC --> FE
    J --> FE
    FE --> S
    FE --> SK
    FE --> C
    FE --> A
    FE --> D
    S --> Final
    SK --> Final
    C --> Final
    A --> Final
    D --> Final
Loading

The engine employs a "dealbreaker collapse" mechanism: if a job fails a hard filter (e.g., salary below comp_floor), the entire score collapses to zero, regardless of other strengths.

For details on the two-track seniority model and the scoring algorithms, see Fit Scoring Engine.

Sources: src-tauri/src/profile.rs#15-34src-tauri/src/pipeline/filter.rs#1-5

Clone this wiki locally