Skip to content

Target Criteria & Fit Weights

Jump to bottom
Chazona Baum edited this page Jun 24, 2026 · 1 revision

Relevant source files

The TargetCriteria system defines the user's career preferences and technical requirements. These criteria act as the configuration for the entire application, driving the high-recall Discovery Prefilter and the high-precision Fit Scoring Engine. All data is persisted in the vault at profile/target_criteria.mdsrc-tauri/src/profile.rs#1-3

Data Model: TargetCriteria & FitWeights

The TargetCriteria struct encapsulates all dimensions of the user's job search, from title matching to compensation floors.

TargetCriteria Struct

Defined in src-tauri/src/profile.rs, this struct is populated by parsing the YAML frontmatter of the profile note src-tauri/src/profile.rs#37-56

Field Type Description
match_titles Vec<String> Recall-oriented aliases used for the initial discovery filter.
target_titles Vec<String> Human-readable titles preferred for display and alignment.
work_arrangements Vec<String> Acceptable modes (e.g., ["remote", "hybrid"]).
target_levels Vec<String> Targeted seniority (e.g., ["senior", "dept-head"]).
comp_floor Option<i64> The minimum acceptable salary.
fit_weights FitWeights Tuning parameters for the scoring engine.

FitWeights Struct

The FitWeights struct contains five integer percentages that must sum to 100. These determine how much influence each dimension has on the final 0–100 fit score src-tauri/src/profile.rs#16-22

Dimension Default Weight Description
seniority 20 Distance between job level and targeted levels.
skills 25 Coverage of required and preferred competencies.
comp 30 Proximity of job high-end to user's floor/target.
arrangement 15 Match between job's mode and user's preferences.
domain 10 Alignment with preferred or avoided industries.

Sources:src-tauri/src/profile.rs#11-34src-tauri/src/profile.rs#37-56

Implementation & Data Flow

The system transitions from "Natural Language" (the Markdown note) to "Code Entities" (the Structs) to drive the pipeline.

Diagram: Profile Data Flow

This diagram illustrates how read_target_criteria transforms vault data into active pipeline configuration.

flowchart TD
    subgraph Subsystems
        PREFILTER["pipeline::filter::prefilter"]
        SCORING["fit::score_fit"]
    end
    subgraph subGraph1 ["Code Entity Space (src-tauri/src/profile.rs)"]
        FM["split_frontmatter()"]
        FRONT["struct Front (Deserialize)"]
        TC_STRUCT["struct TargetCriteria"]
        FW_STRUCT["struct FitWeights"]
    end
    subgraph subGraph0 ["Vault (Markdown/YAML)"]
        TC_MD["profile/target_criteria.md"]
    end
    TC_MD --> FM
    FM --> FRONT
    FRONT --> TC_STRUCT
    TC_STRUCT --> FW_STRUCT
    TC_STRUCT --> PREFILTER
    TC_STRUCT --> SCORING
Loading

Sources:src-tauri/src/profile.rs#96-123src-tauri/src/pipeline/filter.rs#31-35src-tauri/src/fit.rs#18-20

The Discovery Prefilter

The prefilter function in src-tauri/src/pipeline/filter.rs is the first consumer of TargetCriteria. It is designed for recall rather than precision.

  1. Title Matching: It performs case-insensitive containment checks. If a job title contains any string in match_titles, it passes src-tauri/src/pipeline/filter.rs#19-23
  2. Intentional Inclusion: Work arrangement is not filtered here. Even if a user wants "remote," a "hybrid" role will pass the prefilter so the user can decide if a compromise is worth actioning src-tauri/src/pipeline/filter.rs#2-5
  3. Deduplication: It drops URLs already present in the vault and collapses duplicates within the current scrape batch src-tauri/src/pipeline/filter.rs#31-43

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

The Scoring Engine (Fit Hub)

The fit.rs module uses TargetCriteria to calculate the FitBreakdown. This is a two-layer process:

1. Hard Filter Layer

Checks for absolute dealbreakers (e.g., compensation below floor). If a Dealbreaker flag fires, the final score is collapsed to 0 src-tauri/src/fit.rs#3-8

2. Soft Scoring Layer

Calculates five sub-scores (0–100) which are then weighted by FitWeights.

  • Seniority Fit: Uses a two-track model (IC vs. Management). It calculates rank distance (e.g., Senior is 1 rank away from Mid) and applies a YOE Reducer if the candidate's total experience is less than the job's minimum requirement src-tauri/src/fit.rs#72-158
  • Comp Fit: Scales the score based on where the job's high-end salary falls between the user's comp_floor and comp_targetsrc-tauri/src/fit.rs#191-200

Diagram: Seniority Rank Distance Logic

How level_track maps strings to a coordinate system for distance calculation.

flowchart TD
    TC["TargetCriteria.target_levels"]
    JOB["Job.seniority_level"]
    subgraph Track_Management ["Track::Management"]
        M3["'vp' (3)"]
        subgraph Track_Ic ["Track::Ic"]
            M0["'front-line-mgmt' (0)"]
            M1["'middle-mgmt' (1)"]
            M2["'dept-head' (2)"]
            L0["'junior' (0)"]
            L1["'mid' (1)"]
            L2["'senior' (2)"]
        end
    end
    TC -.-> JOB
    M0 --> M1
    M1 --> M2
    M2 --> M3
    L0 --> L1
    L1 --> L2
Loading

Sources:src-tauri/src/fit.rs#74-86src-tauri/src/fit.rs#103-140

Experience Envelope Calculation

To drive the seniority_fit YOE reducer, the system calculates the "experience envelope" from profile/experience/*.md.

  • total_months_experience: Instead of summing individual role durations (which would double-count overlaps), it finds the earliest_start and latest_end across all Experience entities to determine the total career span src-tauri/src/experience.rs#96-127
  • Current Roles: If a role is marked is_current, the latest_end is treated as todaysrc-tauri/src/experience.rs#108-117

Sources:src-tauri/src/experience.rs#96-166

Clone this wiki locally