Spaces:
Sleeping
Sleeping
datasciencedojo
commited on
Commit
•
0511ef6
1
Parent(s):
e21cf92
Update agents/prompts.py
Browse files- agents/prompts.py +42 -1
agents/prompts.py
CHANGED
@@ -115,4 +115,45 @@ Must-Haves: {must_have}
|
|
115 |
Resume: {resume}
|
116 |
|
117 |
"""
|
118 |
-
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
115 |
Resume: {resume}
|
116 |
|
117 |
"""
|
118 |
+
)
|
119 |
+
|
120 |
+
|
121 |
+
|
122 |
+
from typing_extensions import TypedDict
|
123 |
+
|
124 |
+
# Update prompt template to match structured response fields
|
125 |
+
prompt_template_new = PromptTemplate.from_template(
|
126 |
+
"""
|
127 |
+
You are an ATS (Applicant Tracking System) agent designed to analyze resumes and job requirements to assess candidate-job fit. Your task is to match the key skills, experiences, and qualifications from the input resume to the requirements outlined in the job description.
|
128 |
+
|
129 |
+
When performing the match, prioritize "Must-Have" skills and qualifications, followed by other weighted criteria:
|
130 |
+
|
131 |
+
Skills and Keywords (out of 40%): Identify critical skills, tools, and technologies in the Must-Haves.
|
132 |
+
Experience ( out of 30%): Compare years of experience, industries, job titles, and responsibilities, focusing on Must-Haves.
|
133 |
+
Education & Certifications (out of 20%): Match the candidate's degrees and certifications with the Must-Haves.
|
134 |
+
Preferred Qualifications (out of 10%): Compare the candidate's qualifications with the Preferred ones. If the candidate lacks in it, lower the Preferred Qualification score.
|
135 |
+
If the candidate lacks most Must-Haves, significantly lower the match score.
|
136 |
+
|
137 |
+
Ensure that overall_match_score is the exact sum of the individual scores provided above.
|
138 |
+
|
139 |
+
Provide a JSON response in the format below:
|
140 |
+
<candidate_name: name of candidate
|
141 |
+
overall_match_score: sum of scores for skills_keywords_score, experience_score, education_certifications_score, and preferred_qualifications_score (Whole Number)
|
142 |
+
skills_keywords_score: Whole Number score for Skills and Keywords (40%).
|
143 |
+
skills_keywords_explanation: explanation string for Skills and Keywords.
|
144 |
+
experience_score: Whole Number score for Experience (30%).
|
145 |
+
experience_explanation: explanation string for Experience.
|
146 |
+
education_certifications_score: Whole Number score for Education & Certifications (20%).
|
147 |
+
education_certifications_explanation: explanation string for Education & Certifications.
|
148 |
+
preferred_qualifications_score: Whole Number score for Preferred Qualifications (10%).
|
149 |
+
preferred_qualifications_explanation: explanation string for Preferred Qualifications.,
|
150 |
+
score_interpretation: <donot mention any numbers here, just Interpretation in words of the overall_match_score and highlight the key points to explain analysis>
|
151 |
+
|
152 |
+
Ensure the response is consistent if the same resume and job description are provided multiple times.
|
153 |
+
|
154 |
+
Job Title: {job_title_text}
|
155 |
+
Preferred Qualification: {job_listing}
|
156 |
+
Must-Haves: {must_have}
|
157 |
+
Resume: {resume}
|
158 |
+
"""
|
159 |
+
)
|