translate it to english, to a recruiter in a profesional tone """ I remember when I was…
- User
translate it to english, to a recruiter in a profesional tone """ I remember when I was developing a recommendation system to recommend job offers to job seeker users (the parts) so in the very begining versions of the system we just started with the knowledge filters using the known features of both parts and computing the semantic similarity matches between the average embedding of the parts' hard skills, in this point the results were no as good as we expect, but we keep iterating and researching the process and we could use better embeddings models like the OpenAI's ADA and then the 3-large, getting better matches; then we increase the content filter computing the semantic similarity between the Job title and the user's preferred role name, and building a weighted avarage score using the two similarities, and also we could improve the knowledge filter extracting more and in a better way the job offer's features since we change the use of Name Entity Recognition of NLP technic to use LLM request """
- Assistant
Of course. Here are a few options, ranging from a direct translation to a more polished, results-oriented version suitable for a resume or interview.
Option 1: Direct & Concise Translation
This version is a direct, professional translation of your original text, perfect for an email or cover letter.
"I developed a recommendation system to match job seekers with relevant job offers. In the initial version, we used a knowledge-based filter, calculating semantic similarity between the embeddings of a candidate's hard skills and those in a job posting. The initial results did not meet our expectations.
Through continuous iteration, we upgraded our embedding models, first to OpenAI's ADA and later to the 3-large model, which significantly improved our match quality. We then enhanced the system by adding a content-based filter that calculated the similarity between the job title and the candidate's preferred role. We combined these two similarity scores into a single weighted average to rank recommendations.
Furthermore, we improved our knowledge filter's feature extraction process by transitioning from a traditional Named Entity Recognition (NER) technique to using Large Language Model (LLM) prompts, which allowed us to extract more comprehensive and accurate features from job descriptions."
Option 2: Polished & Impact-Focused
This version is more polished and focuses on the process and results, which is ideal for an interview or a project summary.
"In a recent project, I was responsible for developing and iterating on a job recommendation system. Our goal was to accurately match candidates with job opportunities.
We began with a foundational knowledge-based filtering system that compared the semantic similarity of hard skills between candidates and job descriptions. Recognizing the need for improvement, we pursued a multi-pronged strategy:
-
Upgraded Embedding Models: We systematically improved our semantic matching capabilities by transitioning from basic embeddings to more powerful models like OpenAI's ADA and, subsequently, the 3-large model. This directly resulted in more relevant matches.
-
Hybrid Scoring System: We introduced a content-based filter, adding a second layer of analysis that compared a candidate's preferred role with the job title. We then engineered a weighted scoring mechanism that combined both skill and role similarity to create a more holistic and accurate recommendation score.
-
Advanced Feature Extraction: We significantly enhanced our data extraction process by replacing a standard Named Entity Recognition (NER) pipeline with a more sophisticated approach using LLM requests. This allowed us to capture a richer and more precise set of features from job postings, further refining the quality of our knowledge-based filter."
Option 3: Resume Bullet Points
This format is perfect for a resume, as it's scannable and highlights key achievements.
- Architected and iteratively improved a job recommendation engine, significantly increasing match accuracy through a multi-layered filtering approach.
- Enhanced semantic relevance by upgrading from initial embedding models to state-of-the-art models, including OpenAI's ADA and 3-large.
- Developed a hybrid recommendation score by combining knowledge-based (hard skills) and content-based (job title vs. preferred role) similarity metrics into a weighted average.
- Modernized the feature extraction pipeline by replacing traditional NLP (NER) techniques with LLM-based queries, resulting in more comprehensive and accurate data from job descriptions.
-
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.